I. WWWURWWWW W L' 3 1293 00858 40 7 MSU * RETURNING MATERIALS: P1ace in book drop to LJBRARJES remove this checkout from .—r__. your record. FINES will , 7 be charged if book is returned after the date stamped be10w. "$53 6 6 ‘15 A SPATIAL ECONOMIC ANALYSIS OF THE INPACT OF REVERSE OSNOSIS FILTRATION ON THE GRADE A NILK NARKET 8? Ann Hurial Flaaing A THESIS Subaittad to Michigan Stata Univaraity in partial fulfillaant of tha raquiraaanta for tha dagraa of NASTER OF SCIENCE Dapartaant of Agricultural Econoaica 1987 Copyright by ANN HURIEL FLEMING 1987 Th0 u :umm 5"! cu Ev3. an Rim. 2 "‘0 om 3“ bu: This "910m my” "In '92 9:2:an "lam 15cm. Rn“ LR. “9 “Nu. “min ABSTRACT A SPATIAL ECONOHIC ANALYSIS OF THE IMPACT OF REVERSE OSHOSIS FILTRATION ON THE GRADE A HILK NARKET by Ann Hurial Planing Tha axpanding ailk aurplua, falling aupport prica, and inainant introduction of productivity booating tachnologiaa hava craatad an anvironuant of incraaaad tanaion within tha U.S. dairy induatry. Tha raault haa baan an intanaifiad affort to idantify naana for improving aconoaic afficiancy and aatabliahing long-run viability. Ona araa of intaraat haa baan bulk raduction of fluid ailk. Thia thaaia focuaaa on tha acononic faaaibility and ragional iapact of Ravaraa Oaaoaia filtration of Grada A nilk. For analyaia, a ahort-run apatial aquilibriun nodal waa apacifiad. Solutiona wara gonaratad undor a range of pricing and policy acanarioa including Claaa I difforantial ranoval and roalignaont, raducad aupport prica, and incraaaad tranaportation coata. Raaulta indicata that fluid ailk prica would fall with tha dagraa of inpact varying by ragion. Aa a whola conauaara gain whila producara loaa, yat, for aoaa ragiona Ravoraa Oanoaia aay halp ainiaiza tha nagativa impact of cartain policy changoa. Thu mm mm: mm. ' Cum: that 1: 15! IU; 5‘3 .X; ACKNOWLEDGHENTS Thia thaaia raaaarch davalopad into a far mora comprahanaiva and challanging proaact than I had originally anviaionod. In thia aanaa, naivata allowad ma to achiava aomathing which I otharwiaa may not hava choaan to attampt. Cartainly I did not approach thia raaaarch alona. Tha fact that it ia now aitting complatad bafora you ia taatimony to tha aupport and guidanca which I racaivad throughout. I can not axpraaa atrongly anough tha valuabla rola which Larry Hamm had am my raaoarch adviaor and major profaaaor. Hia guidanca and contributiona ara raflactad throughout thia thaaia -- and will undoubtadly influanca futura work. Furthar thanka ia gratafully axtandad to Jamaa Oohmka. Jamaa Hilkar, Tad Farria and Robart Brunnar for thair thorough roviaw of varioua drafta and thair valuabla auggaationa on ‘how to battar thia progact. Throughout working on thia raaaarch I hava racaivad immaaaurabla aupport from poopla whom I am privilagad to call frianda. Thair aupport, concarn and admirabla patianca I ahall navar forgot. With tham, tima apant will navar ba trivial. Finally, a warm thankyou to my family. Thair lova haa alwaya aupportad ma aacuraly, allowing ma tha fraodom to chooaa and tha opportunity to achiava. iv CM. 1.2 ”NMBJ TABLE OF CONTENTS LIST OF TABLES ...-.0...00......0.000000000000000... V111 LIST OF FIGURES ...-......Cll...0.000000000000000... “11‘ LIST OF ABBREVIATIONS .............................. xvi CHAPTER ONE: INTRODUCTION ......................... 1 1.1 Tha Induatry Environmant ...................... 2 1.2 Intaraat in Bulk Raducing Tachnologiaa ........ 4 1.3 Nathod of Analyaia ............................ 8 1.4 Study ObJactivaa .............................. 10 1.5 Ovarviaw of tha Study ......................... 10 CHAPTER TWO: RELEVANT DAIRY INDUSTRY CHARACTERISTICS ...................... 13 2.1 Rola of Govarnmant Uithin tha Domaatic Dairy Induatry ................................ 14 2.1.1 Fadaral Nilk Narkating Ordara ............. 15 2.1.2 Prica Support Programa .................... 17 2.2 Pricing Nachaniama ............................ 18 202.1 a-” Prtc. ......OOOIOOOOOOOOOIOIOOOOOIOOIOO 19 2.2.2 Claaaifiad Pricing ........................ 19 2.2.3 Claaa I Diffarantiala ..................... 20 2.2.4 Bland Prica ............................... 22 2.2.5 Ovar-Ordar Prica .......................... 23 2.2.6 Commodity Exchangaa ....................... 23 2.3 Rola of Producar Cooparativaa ................. 24 2.4 Intarnational Trada in Dairy Producta ......... 24 2.5 Narkat Charactariatica of Supply and Damand ... 25 2.6 Ragional Coata of Production and Compatitiva Advantaga ..................................... 26 2.6.1 Eatabliahing Nichigan’a Compatitiva p°.1ti°n ...-......OO......IIOOIOOOOOOOOIOO 26 2.6.2 Coat of Production ........................ 28 2.6.3 Nichigan’a Ralativa Poaition .............. 30 2.7 Raconatitutad Nilk ............................ 33 2.7.1 Nathoda of Concantration .................. 34 2.7.2 Uaa Raatrictiona .......................... 35 V 2.8 San CEAPTER 3.1 Pri 3.2 The 3.3 Key Rev 3.3.1 3.3.2 3.3 Sun CHAPTER 3.1 Sp: 3.1.1 3.1.2 3-2 60: 3.2.1 3.2.2 3.2.3 3.2.3 1-3 504 3.3.1 3.3.2 3.3.3 3.34 3.3.5 2.8 su-..ry ......COOOOCO......OOOIIOOCOOOOIIOIO... 39 CHAPTER THREE: REVERSE OSNOSIS FILTRATION.......... 41 3.1 Principlaa of Ravaraa Oamoaia Filtration ...... 42 3.2 Tha Ravaraa Oamoaia Syatam .................... 44 3.3 Kay Factora Influancing Oparation of Tha Ravaraa Oamoaia Syatam ........................ 46 3.3.1 Tha Nambrana .............................. 48 3.3.2 Factora Influancing Parformanca ........... 50 3.4 Summary ....................................... 55 CHAPTER FOUR: THE NODEL ............................ 56 4.1 Spatial Equilibrium Nodala .................... 56 44.1.1 Raviaw of Salactad Litaratura ............. 57 «4.1.2 Ganaral Spatial Equilibrium Nodal ......... 58 4.2 Ganaralizad Tranaportation Problam ............ 61 «4.2.1 Vactor Sandwich Nathod Solution Algorithm . 61 4.2.2 GTP Equilibrium Conditiona ................ 62 4.2.3 Prico Linkaga Nachaniam ................... 63 4-2.4 Limitationa of tha GTP Program ............ 64 4-3 Nodaling tha Dairy Induatry ................... 66 ‘4..3.1 Pravioua Nodala of tha Dairy Induatry ..... 66 ‘4-3.2 Nodal Spacification ....................... 68 ‘4-3.3 Prica Linkaga Nachaniam ................... 70 341.3,4 Coat of Ravaraa Oamoaia Filtration ........ 71 ‘3-3.5 Tranaportation Costa ...................... 73 a. Short Haul Function .................... 74 b. Long Haul Function ..................... 76 c. Braak Evan Nilaagaa .................... 77 ‘-3.6 Data Raquiramant .......................... 79 a. Ragiona ................................ 79 b. Tha Tanth Ragion ....................... 80 c. Ragional Cantara ....................... 81 d. Langth of Run .......................... 83 a. Baaa Pariod ............................ 83 f. Bounda ................................. 84 g. Pricaa ................................. 84 h. Damand and Supply Schadulaa ............ 85 (I ’i“lPTER FIVE: ANALYSIS OF RESULTS ................. 89 25"Il 25.. Cavaata and Limitationa ....................... 92 3-2 Baaa Run ...................................... 94 g Impact of RO Filtration ....................... 98 3 ~3.1 Applying R0 Filtration at 8.30/cwt ........ 98 3 ~ 3.2 Applying RO Filtration at ..90/cwt ........ 102 S - 3.3 Applying RO Filtration at 81.75/cwt ....... 104 ~ 3.4 Intarpratation of Annual RO Raaulta ....... 107 vi 5.3 50; 5.3.1 5.3.2 5.5 Cut 5.5.1 5.5.2 5.6 311 5.6.1 5.6.2 5.7 Ad: 5.7.1 5.7.2 539 su CHAPTER 5'1 Ba 5.2 Th 5'3 Su 6.3.1 6.3.2 6.33 5.34 1531103 5.4 Saptambar Conditiona .......................... 109 5.4.1 Saptambar Baaa Run ........................ 110 5.4.2 Application of RO Filtration .............. 113 5.5 Cut in CCC Purchaaaa .......................... 116 5.5.1 Applying RO Filtration at ..90/cwt ........ 119 5.5.2 Intarpratation of Raaulta ................. 121 5.6 Altaring tha Claaa I Diffarantiala ............ 122 5.6.1 Ramoval of Claaa I Diffarantiala .......... 123 5.6.2 Raaligning Claaa I Diffarantiala .......... 126 a. Impact of Baaa Run ..................... 126 b. Applying RO Filtration at 0.90/cwt ..... 128 c. Intarpratation of Raaulta .............. 130 5.7 Additional Runa ............................... 131 5.7.1 Incraaaing Tranaportation Coata ........... 131 a. Impact of Baaa Run ..................... 133 g 1:. Applying RO Filtration at 8.90/cwt ..... 135 5.7.2 Altarnativa Supply Elaaticitiaa ........... 137 5.8 Compariaon of Raaulta with Othar Studiaa ...... 140 Sag SU'..ry aaaaaalaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa 144 CHAPTER SIX: SUNNARY AND CONCLUSIONS .............. 146 6-1 Background to Raaaarch Iaaua .................. 147 6'2 Th. ”06.1 aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa 148 5-3 Summary of Raaulta ............................ 150 6.3.1 R0 Filtration ............................. 150 6.3.2 Saptambar Conditiona ...................... 150 6.3.3 Raduction in CCC Purchaaaa ................ 151 2'3-4 Changa in Claaa I Diffarantiala ........... 152 6.4. $5 Incraaaad Tranaportation Coata ............ 153 6 5 oint. of Caution ............................. 154 ' -c°n°1uaiona ................................... 155 GLOSSARY or 723333 ......IOOIIIOOOOI.........OOIOOIIO 157 app anon: A: SUPPLENENTARY 30043710335 70 ' CHAPTER Two ......CCCCCOCCCICOOICC...-O 161 App EuDIx 8: DATA C.......C.......................-I 163 App “”1" C: 330023.. 5263535750 SOLUTIONS 165 BIB LIOGRAPHY ................................C...... 192 vii 13613 2. 73613 2. 73113 3. a . .3513 4 Table 5 731213 5 131513 5 Tabla 2.1 Tabla 2.2 Tabla 4.1 Tabla 4 . 2 Tabla 5. 1 TCbla 5 . 2 Tobi. 5.3 7‘51. 5.4 7.510 5.5 7‘51. 5.6 T‘blc 5.7 LIST OF TABLES Extant of tha Fadaral Nilk Narkat Ordar Program, 1960 to 1985 .............. Coat of Production, Prica and Raturn to Riak Nanagamant for Salactad St.t..' 1982 (.ICWt) .....IOOIOIOOIOOIOOOO Braak Evan Coat Diatancaa for Thraa Altarnativa RO Filtration Coat Lavala, Ona-Uay Nilaa .................... Ragional Supply and Damand Schadula Slopa and Intarcapt Valuaa ............... Compariaon Batwaan Actual and Nodal Ganaratad Supply Prica and Quantity L.V.1. ......O.........COIOOCOOOOCOOOIOII. Narkat Impact Undar 1985 Conditiona with R0 Appliad at 8.30/cwt, (BRO3) ...... Narkat Impact Undar 1985 Narkat Conditiona with R0 Appliad at 8.90/cwt, (Baas) 0....0.00.0.........OCICOCICCIOCOI. Narkat Impact Undar 1985 Narkat Conditiona with R0 Appliad at 81.75/cwt, (BR0175) 0.000.000.0000...-0.000.000.0000. Narkat Impact Undar Saptambar Conditiona with R0 Appliad at ..SO/cwt. (SRO3) ...... Narkat Impact Undar Saptambar Conditiona with R0 Appliad at 8.90/cwt, Import/Export Ouantitiaa and Pricaa for All Ragiona Undar Nuy’a Naw Supply Elaa- ticitiaa with Ragiona 9’a Elaaticity Sat Equal to .05. W (HUY05> Import/Export Ouantitiaa and Pricaa for All Ragiona Undar Nuy'a Naw Supply Elaa- ticitiaa with Ragiona 9'a Elaaticity Sat Equal to .10, ggtgzia pagibua (NUYlO) xi 178 179 180 181 182 183 184 185 186 187 188 Tabla C122 Import/Export Ouantitiaa and Prices for All Ragiona Undar Huy’a Naw Supply Elas- ticitiaa with Ragiona 9’s Elaaticity Set Equal to .15, cegggia garibus (HUYlS) .... 189 xii Flgur3 1 333333 J 119333 2 119333 1193:. Figuro F19m F391m F393m 51911:. 519111. “9111. F1914“ F1931!. Figura F i gura F i gura F i gura F igura 1" igura l"‘«tgura 5' igura F igura F19ura F‘tguz‘a Fight. F 19111:. F‘Sur. 2.1. 2.22 2.33 LIST OF FIGURES Diatinguiahing charactariatica of tha common mathoda of bulk raduction .... 1985 Claaa I diffarantial prica aurfaca in dollara par cwt .............. Coata of production: variabla and fixad, by ragion for 1984 ............... Total coat of production: aalactad atataa, 1982 ..... Tha oamoaia procaaa ..................... Tha ravaraa oamoaia procaaa ............. Standard mambrana modula uaad for r.v.r.. o..°.1. .........OOOOOOOOOOOOO... Tha ravaraa oamoaia ayatam .............. Kay factora influancing oparation of tha R0 filtration ayatam and aoma alamanta which affact tham ......... Effact of total aolida on parmaata flux ......IOCOCOO Effact of faad flow rata on parmaation flux during tha concantration of akim milk ........ Effact of praaaura laval on parmaation flux during akim milk concantration ..... Effact of tamparatura and parcant protain on ralativa viacoaity of tha concantrata ...... Impact of tranafar coata, 00’, on pricaa and trada batwaan two apatially aaparatad markata xiii 21 30 32 43 43 45 45 46 51 52 53 54 60 Figura 4.2 Figura 4.3 Figura 4.4 Figura 4.5 Figura 5.1 Figura 5.2 Figura 5.3 Figura 5.4 Figura 5.5 I"Agura 5.5 lMaura 5.7 F‘guro 5.8 F19Ur. 5.9 519‘“. 3.10 Salactad apatial atudiaa of dairy markating and acanarioa analyzad ........ Braak avan milaaga function undar nor.‘1co.t. ...-......OOOOOOOOOOOOOOOIOO Dalinaation of ragiona uaad in thi. .tUdY ....IOOIOOO......OOOOOOOOOOIOI Stataa ancompaaaad within ragiona, ragional damand and aupply cantara ...... Rafaranca tabla of modal acanario titlaa and daacriptiona ................. Flow diagram of tha incorporation of varioua acanarioa into tha modal ..... Diatributional pattarn undar 1985 markat conditiona, (BASE) ............... Diatributional pattarn undar 1985 markat conditiona with R0 appliad .t .030/cwt' (BR03) ......OOOOOOOIIOOOOOO Diatributional pattarn undar 1985 markat conditiona with R0 appliad .t ..90/cwt' (BROQ) ...-....IIOIOOOIOIOOO Diatributional pattarn undar 1985 markat conditiona with R0 appliad .t .1a750/Cth (BR0175) OIOIOIIOIOIIOOIOO Diatributional pattarn undar Saptambar markat conditiona, no R0.pp1£.dp (SEPT) .....IIIIOOIOOOOOOOOOO Diatributional pattarn undar Saptambar markat conditiona with R0 appliad at 8.90/cwt, (SR09) .......... Diatributional pattarn undar 1985 markat conditiona with CCC purchaaaa raducad (CCC) .............. Diatributional pattarn undar raducad CCC purchaaaa markat with R0 appliad at 8.90/cwt, Diatributional pattarn undar markat conditiona with Claaa diffarantiala ramovad, (ND) xiv 1985, conditiona (CCCR09).... 1985 I 68 78 80 82 90 91 97 100 102 105 111 115 119 120 123 Figura Figatara F14;\ara Fulfiltara FT‘SI‘Jra l§1LS§ura p ‘1 gura Diatributional pattarn undar 1985 markat conditiona and 1986 Claaa I diffarantiala, (D86) .................... Diatributional pattarn undar 1985 markat conditiona and 1986 Claaa I diffarantiala with R0 appliad at 8.90/cwt, (D86R09) ...................... Braak avan milaaga of R0 filtration undar unadJuatad and fifty parcant incraaaad tranaportation coata .......... Diatributional pattarn undar 1985 markat conditiona and fifty parcant incraaaad tranaportation coata, (TC2) ... Diatributional pattarn undar 1985 markat conditiona with tranaportation coata incraaaad 50 x and R0 filtration appliad at 8.90/cwt, (TC2RO9) ........... Dalinaation of modal ragiona ............. Fadaral milk markating ordar (FNNO’a) and atataa ancompaaaad within ragiona, ragional damand and ragional aupply cantara XV 127 128 132 133 136 166 167 ans ccc (:09 Cut LIST OF ABBREVIATIONS Agricultural Narkating Sarvica Commodity Cradit Corporation Coat of Production Hundradwaight Dairy Narkat Statiatica Dairy Narkat Policy Simulator Economic Raaaarch Firm Entarpriaa Data Syatam Fadaral Nilk Ordar Narkat Statiatica Fadaral Nilk Narkating Ordara Ganaralizad Tranaportation Problam Ninnaaota-Wiaconain National Agricultural Statiatica Sarvica National Nilk Producara Fadaration Ravaraa Oamoaia Ultrafiltration Ultra High Tamparatura Unitad Stataa Dapartmant of Agricultura Unitad Stataa Dapartmant of Commarca Unitad Stataa Public Haalth Sarvica Vactor Sandwich Nathod xvi CHAPTER ONE INTRODUCTION Govarnmant afforta to curb milk aupply and raduca QOVanmant axpandituraa within tha dairy induatry ara £°rcing tha induatry towarda improvad aconomic afficiancy. producara, finding that thay muat raduca axpanaaa in °rdar to ramain viabla, ara looking for maana by which to Ona araa “1ntain ravanuaa and aacura thair livalihooda. Q 2‘ obvioua potantial for coat raduction ia tha tranapor- t ‘tion of bulk fluid milk. In a 1980 National Economica b"‘Iiaion ataff raport. Ed Jaaaa datarminad that aubatantial ‘§Qnomic incantivaa axiat in favor of concantrating and hkfin raconatituting whola milk whan tha diatanca batwaan b*‘Oduction and conaumption pointa axcaada 100 milaa. Latar ‘\\Jd1.l by Novakovic and Aplin (1981), Novakovic (1982), ‘hd Whippla (1933) hava aubatantiatad tha potantial ‘§onomic incantivaa in ahipping milk in a concantratad %§1m. Although thara axiat a multituda of mathoda for ‘b‘aoving watar from milk , racant tachnological advancamanta “ ‘3- ravaraa oamoaia (R0) filtration, a mambrana filtration \‘chniqua, hava ganaratad intaraat by producar organ- :‘ ‘ationa. Ona advantaga of R0 filtration ovar othar 0.3 g 3 53. 3’6! 153 '1 3 I/ 2 tachnologiaa ia that it producaa a auparior product holding considarabla promiaa for conaumar accaptanca. Thia thaaia ia daaignad to datarmina ravaraa oamoaia filtration'a potantial for adoption by tha Claaa I (fluid 3.3a-) milk markat undar a nonraatrictiva policy anviron- tht, and to datarmina what tha impact of much adoption Would ba on a ragional and induatry-wida baaia. It ia b.11avad that R0 filtration of fluid milk could provida producara in tha Uppar Nidwaat with tha opportunity to e‘.E>1taliza on thair compatitiva advantaga in milk produc- tion, by capturing naw markata at graatar diatancaa. Tha b“ulta obtainad from thia raaaarch prova to ba both For j‘hCightful and, in aoma caaaa, rathar aurpriaing. ‘T‘Qmpla, no production ragion ia found to banafit from R0 3 ‘~ ltration to tha dagraa initially hypothaaizad. Tha ramaindar of thia chaptar focuaaa on tha currant ‘hwironmant undar which tha induatry ia oparating, raaaona t§=33~ intaraat in bulk raducing tachnologiaa, and tha mathod ‘3 analyaia utilizad in datarmining tha impact of R0 5 i ltration upon tha markat. Thia will ba followad by a ‘tatamant of raaaarch obJactivaa and an ovarviaw of tha §haptara compriaing thia thaaia. ‘ ~— 1 Tha Induatry Environmant Undar tha praaant aconomic, tachnological and lagia- l Qtiva anvironmant, tha dairy induatry facaa an ara of ‘fignificant changa. Racant raductiona in tha dairy aupport 3 prica hava forcad many producara to drop out of production whila othara hava anliatad in tha govarnmant aponaorad Dairy Tarmination Program. Naanwhila. amarging tach- nologiaa promiaa to axpand production lavala to naw haighta. For axampla, it haa baan aatimatad that bovina lo‘atotropinl alona could incraaaa long-run productivity by .Pproximataly 15 parcant. Such a ahift foratalla of a ‘.J°r raahuffling within tha induatry aa inafficiant pro- duCQra ara forcad out and thoaa ramaining in production -xpand thair oparationa in ordar to captura tha banafita of -Qfila aconomiaa. Tha affact of thaaa two oppoaing forcaa, th- naad to dacraaaa aupply and tha rapid incraaaa in par unit production, haa craatad an atmoaphara of incraaaing 1: ‘hsion within tha induatry. Claarly naw tachnologiaa will hava a kay rola in datar- {hing tha diraction and compoaition of tha induatry Q ‘ar tha naxt dacada. Juat aa aoma tachnologiaa may impact ha induatry through incraaaing aupply, othara may halp Spacifically, : §‘<:>clucara through thia tranaitional pariod. ‘3 1k raduction through mambrana filtration tachniquaa haa §‘naratad conaidarabla attantion. Thia procaaa could ET§tantially allow producara to banafit according to thair =§:l3patitiva advantaga in production whila tha induatry ‘Qvancaa .towarda improvad aconomic afficiancy. \ 1 Bovina aomatotropin, commonly rafarad to aa bovina 9“owth homona, ia a naturally producad hormona within tha himal which. at incraaaad lavala. aignificantly anhancaa “ 1 1k productivity. 1|: 'V. 213 4 1.2 Intaraat in Bulk Raducing Tachnologiaa Fluid milk containa approximataly 87 parcant watar (Naaraon and Ginnatta, 1970). Thia watar contant, dua to its bulk and waight, incraaaaa tranaportation. handling. atoraga, and praaarvation coata. Claarly than, a natural air... for coat raduction within dairy markating ia through erucing tha milk’a bulk prior to ahipping. Tha praaant markating ordar atructura aupporta thia cciricluaion. Changaa in milk markating ordara hava in- erOQaad tha diatanca which milk ia movad within ordar t5.910na, with diatancaa oftan axcaading 1,500 milaa (Jaaaa, 13BO). Givan thia ordar tranafar atructura, it aaama t‘haly to raappraiaa tha currant tranaportation ayatam. Oh. aolution to raaolving tha tranaportation coat problam “Quid ba to tranaport a concantratad product, which can {kan ba "racombinad" into a whola fluid product naarar tha b“ int of conaumption. Tha aconomic advantagaa of ramoving ‘§ma of thia watar can ba actualizad only aa long aa tha §Qat of auch ramoval doaa not outwaigh tha aavinga acquirad through raducad bulk and waight. In tha U.S., tha majority of ragiona produca an adaquata =€ luid milk aupply to maat ragional damand. Importing fluid ‘ ilk from othar ragiona may ba nacaaaary during timaa of k‘mporary or aaaaonal ahortaga. Givan tha viability of §E‘conatitutad milk am an altarnativa or aupplamant to lQcal milk production in daficit ragiona, ona would axpact kw aaa a raduction in gaographic milk prica diffarancaa 5 among ragiona. Tha raaaon for thia ia that with Claaa I diffarantiala baaad on whola milk tranaportation coata and tranaportation coata baing poaitivaly corralatad with 9.19111: and voluma, ramoving tha watar dacraaaaa tha coat of ahipping milk. l’or ragiona auch aa Nichigan which produca a high POrcantaga of fluid grada milk (97 parcant of total produc- tiOn) and which hava a claar compatitiva advantaga in prOduction ovar all ragiona axcluding tha Uppar Nidwaat, th- potantial impact of raconatitutad milk ara particularly -ttractiva (NNPF, 1985). Ralativaly low coat of production t.‘910na could aaa a riaa in axport damand laading to an :‘hcraaaa in tha proportion of production going to Claaa I “‘a. Am a raault, producar ravanuaa could incraaaa. §‘rtainly tha potantial banafita of raducing tha bulk and “‘ight of milk appaar attractiva but tha mathod of concan- ~§§‘ation ramaina important. aapacially in tarma of conaumar ‘ Qcaptanca . Traditionally, avaporation and apray drying hava baan ‘§ployad to ramova watar from akim milk and mora racantly \ha ultra filtration procaaa haa gainad attantion. Liquid :sQoda, howavar, ara vary vulnarabla to flavor and aroma §laangaa and aach of thaaa procaaaaa hava nagativa aida ‘d'facta. For axampla, axpoaura to high tamparaturaa undar ‘pray drying can altar milk'a charactariatica and ultra t dltration may raault in tha loaa of aoma nutrianta praaant 3~ 23 raw milk. 6 Hora racantly, ravaraa oamoaia filtration haa baan appliad to tha concantration of fluid milk. Tha advantaga of R0 filtration ovar tha mora traditional forma of raduc- tion liaa in ita nondapandanca on haat. A larga parcantaga Of watar can, tharafora, ba ramovad without altaring taata find nutriant charactariatica: thua, minimizing tha impact on conaumar damand. It ia for thia lattar raaaon that tha R0 filtration procaaa haa gainad graat intaraat among bulk r-duction tachnologiaa. With tha aid of tachnological -dvancamanta, it appaarm tha procaaa ia alao bacoming an -§°nomically faaaibla altarnativa. A briaf ovarviaw of -°‘a common bulk raducing altarnativaa ia praaantad in s‘:.~Sura 1.1. Claarly, amploying RO filtration praaanta aoma vary ‘*citing poaaibilitiaa: (1) low coat production ragiona §§\ildzl banafit from thair compatitiva advantaga: (2) tha { hduatry could raalign ita production and ahipmant b"ttarna, gaining incraaaad aconomic afficiancy: and, (3) §§naumara could banafit from tha raduction in tranapor- t‘tion coata aa raflactad in tha ratail prica of milk. t‘gionally and intraragionally tha impacta would cartainly T ‘ary. For axampla, conaumara in ragiona farthaat from tha ‘fiurca would nacaaaarily aaa tha graataat potantial “ hpact. Furtharmora, low incoma conaumara. who davota Q. graatar proportion of thair incoma to food purchaaaa. §Quid raaliza appraciabla banafita. .Nnaa ..~u .u muzrdoudom van and“ .oonuon "uuuaom domuuavou a—oa «a upanuua 803809 on. no .umuumuouunuwnu mama-«nudmuuumov.n.n own-mm domuaumundouou can udmuuuuoud .5538»... «o .aoU o «doadmaco aomuwauoduduuu nudmuuxuwa aoma-«mu Lo «.09 cumin a_ma uaad: moan: nodauodOi >x~an3 nuuuu «no: 6 m.on cam—onnm 3.06.0 «woo udoaamavu Audra domuuuudoudOu uuuo>um nouMaVou can.» udomuuad a. x—«E Ema. pa. \dom.uu.~mm one: can uuo_ «So. was mau domuuopou song. on +.~n «nu—D on... aomuoudua aoma-mu». domuwuuaoudau nuauwnu uuou—w «a cannon uuudaumumdummu noptod uuuuaouu o>mndoudm Mama Ema. udmuuo on. no>muAuw3 uoau panama: mamuauoum mu a.o~« madam domuaumuudouou aaa puvood «on oum— «uptown «aaama Jame RID ud- udmuouuauacwa domuauudoudou . woundedw>o uou auburn \domu «a oouaou nonmamuu domuwuuaoudou oomuauoau>m -32... 6.26- 2333.. 323.2,..- 2:... 225 3. 3 on a... 23.3... udmmup a» nomad admsup «no: numunmuouuauaAu x_ma Ema. domuuu0du>m no domuade—ou unqumbau- atodaaDu admuuovdou RN a.no~ _ndom.ou>cou Auvm_om $5 “u3u\.uauuv E Eggs a: SEQ $23528 58 Si 8 Thia wide range of potential benefita froa the adoption of R0 filtration technology clearly providee aufficient merit to atudy ita introduction into the fluid nilk induatry. However. it ia recognized that not all tech- nologiee will be beneficial to all producara. Given thia, it beconea inperative to deteraine the proaected inpact of a given technological change at both an induatry and regional level. 1.3 lethod of Analyaia There are a wide range of econonic iaauea aaaociated with the adoption of R0 filtration. 0f apecific intaraat to thia atudy ia R0 filtration'a affect on fluid nilk aarketa. The tranaahipeent and regional nature of thia gueation euggeata the need to apatially nodal the national fluid nilk narket. The baaic theory behind apatial price equilibriun illuatratea why auch analyaia ie uaeful. Spatial theory of pricing euggeate that in the abaence of narket diatortiona, cauaed by adniniatered pricing or monopoly preaence. the difference in price between geo- graphic ragiona will be equivalent to the coat of moving the product between regiona. Hence, in a trading environnent free of barriera, ea long aa the aupply and denand equilibriun price difference between the two ragiona ia not leaa than the coat of tranaportation, trade will occur. Milk will be ahipped fron the eurplua (relatively low price) to the deficit (relatively high price) region. 9 The conaequence of a decreaae in tranaportation coat and, hence, a change in the regional price wedge, would be evidenced by an altering of the pricea, quantitiea traded, and diatribution patterna within the 0.5. fluid milk market. For example, ragiona which face very high fluid milk pricea would expect to aee a decreaae in price and an increaae in conaumption. Furthermore, if production coata are relatively higher in the deficit region, a decreaae in production could be experienced. On the other hand, a aurplua and relatively low coat of production region, like Michigan, might enjoy many benefita ariaing from R0 filtration and reconatitution. Export markata may expand and the blend price increaae aa aurplua fluid grade milk, which would otherwiae be ”dumped" into leaa prof- itable manufactured dairy product uae, ia allocated to Claaa I uae. Spatial equilibrium analyaia will allow determination of the economic and diatributional impacta of introducing R0 filtration to the Grade A market. Several apatial modela of the dairy induatry have been developed (Hallberg et al., 1978: and Novakovic et al., 1980). Although theae modela are very thorough, they create a limitation aimply becauae of their aize -- large mainframe modela with extenaive data requirementa. For the purpoaea of thia atudy, a general idea of the market impact of R0 filtration can be obtained through the uae of a more aimpliatic apatial equilibrium model run on a 10 leaa demanding computer program. Sharplea and Holland (1984) and Holland (1985) have developed auch a program, the Generalized Tranaportation Problem (GTP). Although GTP ia apecified for international trade, apatial price theory ramaina the aame for trade between any apatially aeparated ragiona. The program generatea trade flow quantitiea, pricea and revenuea and haa a data requirement commenaurate with the detail level of thia atudy. In aum, GTP and apacial equilibrium theory preaent the neceaaary toola of analyaia for obtaining the reaearch objectivea. 1.4 Study Objectivea The objectivea of thia atudy are: (1) To gain a working knowledge of R0 filtration. con- atrainta to ita adoption. operational parametera and ita potential for future uae within the Claaa 1 milk market. (2) To underatand the fundamental characteriatica of the dairy induatry; to reaffirm the relative production advantagea among atataa and ragiona: and. to identify market relationahipa and policiea relevant to the aale of R0 filtrated milk. (3) To incorporate theae market characteriatica and relationahipa into the deaign and apecification of a apatial equilibrium model of the Grade A milk market. (4) To examine the potential Claaa 1 market impact of the full acale adoption of R0 filtration through analyaia of model reaulta run under a range of pricing and policy acanarioa. 1.5 Overview of the Study Thia atudy evaluatea the economic feaaibility of R0 filtration through apatially modeling a aelected aegment of 11 the 0.8. Grade A milk market. The modeled area ia broken down into nine aupply and ten demand ragiona. The necea- aary aupply and demand data are reapecified from atate and federal marketing order atatiatica. Eatimated coata of applying R0 filtration, tranaportation coat functiona, and aupply and demand elaaticitiea are all extracted from previoua reaearch. A apatial equilibrium model incorpor- ating thia data ia apecified according to market character- iatica and the objectivea of thia theaia. Raaulta from thia model are generated under numeroua economic acanarioa. The regional impact on diatribution, pricea and revenuea are iaolated for croaa analyaia. Thia theaia ia divided into aix chaptera yielding inaight into the operation of R0 filtration, the induatry in which ita adoption ia being analyzed, and ita potential impact on fluid milk markata. Chapter Two reviewa fundamental characteriatica of the domeatic dairy induatry. Areaa covered range from the role of government, the pricing mechaniama involved and regional coata of production. Diacuaaion focuaaa on induatry relationahipa which are tied to the adoption of R0 filtra- tion. Additionally, the iaauea related to raconatitutad milk are diacuaaed. Chapter Three deala excluaively with the technology iaolated for atudy: R0 filtration. The general principlea of R0 filtration are preaented and the R0 ayatam itaelf ia described of R0 for In Che zntroduce aspatia] equilxbri veil as 1 11h: up: Iodel uu data req‘ Model Fm. n ”9 due: Iodel, associat‘ Chapt. Obtain“ " "911 A "Wits, l2 deacribed. Potential operational conatrainta to adoption of R0 for fluid milk are highlighted. In Chapter Pour the relevant induatry characteriatica introduced in Chapter Two are placed within the context of a apatial equilibrium model. The general theory of apacial equilibrium and previoua modeling reaearch are diacuaaod as well aa the apecific computer program and aolution algor- ithm employed in thia atudy. Finally, the fully apecified model uaad for analyaia ia aubmitted along with ita data requirement. Model reaulta and analyaia are praaantad in Chapter Five. The acanarioa under which each aolution ia generated are deacribed, aa well aa their incorporation into the model. Additionally, the primary limitationa and caveata aaaociated with the model are aubmitted. Chapter Six aummarizea the atudy and the reaulta obtained. The achievement of atudy objectivea ia diacuaaed aa well aa what can and can not be inferred from the reaulta. In concluaion, implicationa of the reaulta are aubmitted. ha ti CHAPTER 2 RELEVANT DAIRY INDUSTRY CHARACTERISTICS The dairy induatry, by ita nature, poaea unique markat- ing concerna. Fluid milk ia a bulky, highly pariahable, continuoua flow product aubJect to health contaminanta: requiring atrict aanitary compliance in production, trana- portation and proceaaing. Furthermore, aupply ia highly inelaatic in the ahort-run with producara traditionally being very vulnerable to the market power held by pro- prietary handlera and proceaaora. Theae characteriatica all poae potential marketing problema. Hilk alao haa important nutritional qualitiea. Notably it ia high in protein and calcium and ia conaidered an important nutritional item in the nation’a diet. It is becauae of thia combination of aenaitive marketing condi- tiona and nutritional importance that the dairy induatry haa been aeparated from other agricultural induatriea in the deaign of marketing policiea. Dairy induatry policy ia unique in that dairy ia the only induatry where both government price aupporta and federal marketing ordera axiat. The degree of regulation and the complexity of the pricing mechaniama which have evolved, together with the 13 14 diatinguiahing marketing problema aaaociated with milk, make it difficult to tranafar general knowledge of agricul- tural marketing and policy to the apecific concerna of the dairy induatry. Hence, before one can approach the taak of apatially modeling the marketing of fluid milk within the United Stataa, a baaic underatanding of the induatry ia eaaential. Thia chapter overviewa aome of the more prominent dairy induatry characteriatica, focuaing on areaa directly tied to thia atudy. An examination of the regional production coat atructure upon which the relevance of thia atudy hingea will follow. Finally, a cloae look at reconatituted milk will be made, including mathoda of concentration, the role of current regulationa, and aomo previoua atudiea addreaaing theae iaauea. 2.1 Role of Government Within the Domeatic Dairy Induatry The government haa intervened in fluid milk markata by creating regulationa to reduce inequitiea, uncertainty, and variability. In general, the government intervenea in commodity markata becauae there are unaatiafactory condi- tiona within the market. For example, milk’a periahability createa the opportunity for groaa inequitiea to develop, reaulting in market uncertainty and price and quantity inatability within the induatry. Before regulation, proceaaora could ahift coata between producara, manufacturea and fluid planta, all of which had supply 1 lid to inst. mkoting p dmlopod ( ubzlxty, t opmton ‘ “Mn, 04 clmly cl Addmo mm to Maury and mm Such luv“ “a Grade ”may. MC! “.1 muting ‘ Mien" ‘ of “den 15 had aupply level requirementa to enaure efficiency. Thia led to inatability in the general milk market and, with marketing power in the handa of proceaaora, inequitiea developed (USDA, Jan. 1984). Furthermore, to limit var- iability, reaervea needed to be maintained. Private plant operatora were unwilling to bear the burden of thia expenae, aa were producara. The fluid milk market waa clearly claaaifiable ea diaorderly. Additionally, milk auppliea often became contaminated or failed to meet health atandarda. Incentivea were deemed neceaaary to encourage inveatment in the coatly equipment and fecilitiea required to improve aanitary atandarda. Such inveatmenta would enaure that an adequate aupply of aafe Grade A milk waa alwaya available. Hence, the central objective of many dairy program proviaiona ia to provide price atability and an equitable income to producara, while enauring a reliable and aafe aupply of milk for the nation’a conaumara. Regulation primarily came in the form of federal marketing ordera and price aupport programa. 2.1.1 Federal Iilk Marketing Ordara Federal regulation of fluid milk markata began with the eatabliahment of the Federal Milk Marketing Order (FHHO) ayatam in the 1930'a. Thia ayatam waa daaignad to addreaa the chronic problema exiating within the market. Federal ordara for milk are iaaued by the Secretary of Agriculture and adminiatered by the Dairy Diviaion of the USDA'a “ “vim...“ . Agricu speclf it! OI fruit funct: przcx' indir lllk the too IQt {Q1 16 Agricultural Harketing Service. Each order appliea to a apecific geographic region where producara have voted for ita eatabliahment. The function of FHHOa differa from thoaa common to fruit and vegetablea in that milk marketing ordara actually function to eatabliah an inatitutional atructure for pricing. FHHOa regulate all the fluid milk induatry indirectly via direct regulation of handlera aelling their milk within ordara. Thia ia accompliahed by aetting a minimum price which muat be paid by proceaaora to producara for Grade A milk1. Proceaaora may then uae the milk for any purpoae, including manufactured producta which only require the lower quality atandard Grade 8 milk. In the 1930’a, when the FHHOa were eatabliahed, markata were local. Seldom waa milk tranaported between markata; the technology waa not advanced enough and the riaka were too high. Aa a raault, aupply and demand were neceaaarily met within the market. Due to tranaportation coata being greater for fluid than for manufactured dairy producta, aurplua ragiona developed large manufactured product induatriea. Theae producta were marketed on a national acale: fluid markata remained relatively local, within a couple hundred milaa. Aa the technology (tranaportation and refrigeration) improved, tranaportation coata and riak decreaaed. Aa a raault, the 1 Only Grade A milk ia regulated under federal ordara. D01 811' th un CC 17 neceaaity for adequate reaervea to be held locally declined and federal ordara began to merge. In 1962 the number of FHHOa peaked at 83. Since then the number haa decreaaed, yet, the amount of milk covered under federal ordara haa increaaed. At praaant, approxi- mately two-thirda of all milk marketed in the 0.8. ia covered under federal ordara. Theae trenda are illua- trated in Table 2.1. Table 2.1. Extent of The Federal Hilk Market Order Program, 1960 to 1985 Volume of Volume aa a Number of hilk Covered Percent of Year Federal Ordara Under Ordara 0.3. total (Number) (Bil. lba.) (Percent) 1960 80 48.8 45.0 1965 73 54.4 48.3 1970 62 65.1 59.6 1975 56 69.2 62.8 1980 47 84.0 67.4 1985 44 97.8 70.0 Source: National Nilk Producara Federation, 12§1_Qgigz, W. 1985. p-20 and USDA. W W- 1986- 2.1.2 Price Support Programa Current government policy intervention in the dairy induatry ia primarily in the form of price aupporta which were eatabliahed in 1949. Acting through it’a Commodity Credit Corporation (CCC), the government guaranteea purchaae of chaeae, butter, and nonfat dry milk at a aat price. Theae CCC purchaaea aerve to aupport the market price of manufactured dairy producta and indirectly help to 18 aupport the price of all milkz. While the price aupport program ia not directly tied to the federal order program, it doea have a direct impact on federal order marketinga. 2.2 Pricing Rechaniama Price ia the primary coordinator of activity at each atage of marketing: production, aaaembly, proceaaing and diatribution. Price alao aervea aa a production incentive and helpa to maintain adequate auppliea for the varioua competing aourcea of demand. Pricaa for milk and dairy producta are partly adminiatered and partly negotiated in the market place. The current price atructure ia a raault of the combined influence of government regulationa and cooperative action aa allowed by the Capper-Volatead act. Government regula- tiona directly impact the u-w price through price aupporta and impact the Claaa I milk price and the blend price3 through claaaified pricing. Producer cooperativea have created over-order premiuma, now common in moat ordara. Each of theae pricing mechaniama ia diacuaaed below with their mathematical formulationa preaented in Appendix A. 2 While Claaa I milk receivea a price equal to the u-w (Grade 8 milk) price plua a local differential, government purchaaea aupporting the manufactured gooda market price indirectly aupporta the fluid grade market. 3 The blend price ia that which the producer receivea and ia determined by proportion of milk allocated to the alternative Grade A uae claaaea. 19 2.2.1 l-H Price The Hinneaota-Uiaconain Price Seriea ia a weighted average price paid for non-order (Grade 8) milk deatined for manufactured uae in the atataa of Hinneaota and wiaconain. Computed on a monthly baaia by the USDA, it ia deaigned to reflect a combination of the wholeaale product price level and manufacture'a profit margin. The former may be partially determined by aupport pricea and the latter ia determined in the market place. Aa a raault of thia atructure, under FIHOa the price of milk deatined for atoreable manufactured producta ia aet equal to the H-W price. 2.2.2 Claaaified Pricing Federal milk marketing ordara regulate via eatabliahing a minimum price which handlera muat pay and which producara receive for Grade A raw milk. Although handlera buy milk from producara, the price they pay and price producara receive ia not the aame. Specifically, handlera pay what are termed claaaified pricea. Claaaified pricea are baaad on the end uae, or Claaa, to which the milk ia put. Depending upon the order, there are either two of three claaaificationa: Claaa I producta compriae freah fluid producta; Claaa II are aoft manufac- tured producta: and Claaa III are hard manufactured producta. Where there ia no Claaa III diviaion, Claaa II repreaenta all manufactured producta. 20 The Claaa I price atructure differa greatly from that of Claaa II milk. Claaa I milk receivea a higher price than doea milk going to either Claaa II or III uae. Thia ia partly neceaaitated by the Claaa I product’a nature: more expenaive to tranaport and more vulnerable to apoilage. Regionally, the Claaa I price will vary according to the Claaa I differential, aa diacuaaed in the following aection. In contraat to the Claaa I price are the Claaa II and III price which the USDA indirectly aupporta through the u-w price. Specifically, a tentative Claaa II price ia announced for the following month baaed on a formula. Thia formulation takea the l-U price for the aecond preceding month and adJuata it via the weighted change in the groaa value of milk uaed to make cheddar cheeae and butter/nonfat dry milk. Hence, the Claaa II price ia nearly equivalent to the u-w price and ia fairly uniform nationally, while Claaa I pricea vary poaitively with diatanca from the Upper Hidweat -- the traditional milkahed. Thia pricing theory ia reflected in the obaerved differential pricing acheme. 2.2.3 Claaa I Diffarantiala Federal order Claaa I milk pricea are aligned to Eau Claire, Uiaconain, the milkahed'a baae point. The Claaa I differential increaaea directly with incraaaing diatanca from Eau Claire in auch a way aa to approximate the coat of tranaporting raw milk from the milkahed. Figure 2.1 illuatratea thia apatial pricing ayatam: the higheat F19“: d°ll¢ 21 Figure 2.1 1985 Claaa I differential price aurfaca in dollara per cwt.‘ Source: Data taken from FHOS, USDA; Contoura, aatimatad. 4 Note, thia price aurfaca ia an approximation derived from Claaa I price differential data for aalactad citiea. While thia figure doea provide an illuatration of the general caae, not all price differentiala coincide with thia aurfaca. 22 differentiala are in Florida and the loweat in hinneaota. Theae differentiala are tacked onto the baae price which handlera muat pay for milk going to Claaa I uae. Hiator- ically, the USDA haa met the Claaa I differentiala. With the paaaage of the 1985 Food Security Act, thia pricing role waa taken over by Congreaa. Although Claaa I differentiala reflect tranaportation coata of raw milk, they are not the primary determinant for whether milk will be tranaported. The price which pro- ducara receive, the blend price, aervea that function. 2.2.4 Blend Price While handlera muat pay for Grade A milk according to ita end uae, producara do not receive payment according to how their milk ia uaed. Rather, producara receive what ia termed a blend price for their milk. To preaerve equity among producara, regardleaa of their diatanca from the fluid milk market, proceeda from both fluid and manufactured aalea in a given order are pooled. The proceeda are than diatributed to producara at a blended per unit price, with allowancea made for location, butter- fat and marketing aervicea5. Hence, the blend price ia 5 For example, if 80 percent of the milk mold in an order goea to Claaa I producta and the remaining 20 percent to Claaa II, than the blend price per unit received by all producara aelling in that order ia compriaed of 80 percent of the Claaa I price and 20 percent of the Claaa II price. 23 determined by the proportion of milk pooled on a given order going to each Claaa uaage. 2.2.5 Over-Order Pricing One conaequence of cooperative growth within the federal order ayatam haa been the development of over-order pricing practicea. Over-order premiuma repraaant an additional charge inatituted by a producer cooperative which the handler muat pay. While cooperativea announce the over- order premium for the market in advance, decreaaing the uncertainty faced by handlera, theae pramiuma generally have not been negotiated. It ahould be noted that pro- ducara may not receive the full over-order premium. Producer cooperativea often take a proportion of the premium to cover operating expenaea. Both the amount of the premium and the amount which the cooperativea withhold variea between ragiona and over time. 2.2.6 Commodity Exchangaa For two manufactured gooda, butter and cheeae, formal commodity exchangea have been eatabliahed at the Chicago Mercantile Exchange and the National Cheeae Exchange, reapectively. A primary aervice generated by theae exchangea ia that they form the baaia for formula pricing, location price adJuatmenta, and product characteriatic adJuatmenta. 24 2.3 Role of Producer Cooperativea The initiation of federal order regulationa effectively decreaaed the power of proceaaora while incraaaing atabil- ity in the market. One conaequence of theae regulationa haa been a ahift in market power from induatry proceaaora to producer cooperativea. Theae cooperativea began appearing in the late 1960'a and have a growing role within the induatry. Cooperativea act to procure, aaaemble and coordinate the cyclically contraating aupply and demand. Addition- ally, they may provide aervicea auch aa quality control, intermarket tranafera and aurplua management. Preaently more than 85 percent of producara in federal ordera are cooperative membera and more than 75 percent of the nation'a milk, ia aold through cooperativea (USDA, Jan. 1984). The primary direct impact which cooperative preaence will have on thia atudy ia through the exiatence of over-order premiuma. 2.4 International Trade in Dairy Producta Dairy ia a highly regulated and protected induatry in moat modern induatrialized countriea. Domeatic dairy programa commonly have led to aignificant aurpluaea and, in turn, to the impoaition of import barriera and/or heavy axport aubaidiea. In the U.S., importa of dairy producta have averaged leaa than two percent of total U.S. milk production annually (USDA, Jan. 1984). On a world wide 25 acale, trade in dairy producta ramaina fairly ateady near five percent of total world milk production (USDA, Jan. 1984). The world trade in dairy producta will likely remain relatively amall and aa auch, will not affect the analyaia undertaken in thia atudy. 2.5 larket Characteriatica of Supply and Demand Fluid milk markata diaplay unique characteriatica in both aupply and demand. A fundamental characteriatic of the fluid milk market ia that aeaaonal patterna axiat in both production and conaumption: however, theae patterna do not coincide. Production peaka in late apring and trougha in late fall, while conaumption ia loweat in late apring and higheat in early fall. Similarly, conaumption ahowa atrong weekly trenda. Hence, for demand to be met on any given day of the year neceaaitatea aurpluaea at other timea. Of particular intereat to thia atudy are the aupply and demand elaaticitiea for fluid milk. Dairying ia charac- terized by a highly inelaatic ahort-run aupply achedule. Thia ia due to high fixed inveatmenta in apecialized facil- itiea which prevent rapid contraction or expanaion. Additionally, there ia a lag of two yeara from birth until a heifer entera the milking herd. Aa a raault, aupply ia very unreaponaive to price over a two year period, becoming more elaatic in the long-run. 0n the demand aide, dairy product aalea characteriatically are not very price 26 reaponaive in the ahort-run and may be more reaponaive in the long-run. 2.6 Regional Coata of Production and Competitive Advantage The exiatence of competitive advantage in a free market allowa one region to benefit from producing the commodity for which it haa a competitive advantage. The queationa addreaaed in thia atudy are relevant only ao long aa the Upper Midweat really doea have a competitive advantage in the production of milk. 0f apecific intaraat ia Michigan'a poaition relative to atatea outaide of the Upper Midweat region. 2.6.1 Eatabliahing Michigan'a Competitive Poaition Regional competitive poaitiona in the dairy induatry are linked to coata of production. Given thia, it ia neceaaary to compare the coata of producing milk in Michigan relative to other ragiona in order to eatabliah the competitiveneaa of Michigan’a dairy induatry. Average coat of production figurea for milk in Michigan can be obtained from two reliable aourcea, Michigan State Univeraity Telfarm reporta and the USDA’a Firm Enterpriaa Data Syatem (FEDS) budgeta. Telfarm ia a Cooperative Extenaion aupported farm accounting project operated by Michigan State Univeraity. It generatea Michigan farm accounting recorda by uaing a voluntary mail-in ayatam. Several hundred dairy farma take part in thia proaect. 27 Although the data generated may not be repreaentative of all dairy farma in Michigan, given the large aample aize, one can be confident that the data doea accurately repre- aent commercial Grade A dairiea groaaing greater than 850,000 annually. The USDA’a coat of production atudiea were mandated by the federally legialated Agricultural Conaumer Protection Act of 1973. Thia Act requirea annual reporta to Congreaa on the coata of producing varioua commoditiea, including milk. Although data ia originally compiled at the atate level, the USDA publiahea ita coat of production atatiatica annually in regional form. A large part of the FEDS technical data uaad to eatimate coata of production ia compiled by the Economica Reaearch Service (ERS) and National Agricultural Statiatica Service (MASS) through enumerated aurveya of farm operatora. ERS preaenta coat of production data in enterpriae budget form: a liating of all the coata and returna aaaociated with the production of a apecific commodity. State enterpriae budgeta are generated for each atate located in the major production ragiona for the given commodity. Theae budgeta are then weighted according to production, determining the regional and national average production coata for that commodity. Caah receipta are alao weighted in thia manner. Opportunity coata for feeda, unpaid labor, and capital are uaed while coata of machinery and buildinga are generated from a data baae of thoae itema. 28 There are two major areaa of weakneaa aaaociated with the USDA generated data. Firat, the uaual problema aaaociated with aggregating data are likely to exiat (ie. intraregional variation ia maaked). Second, the USDA’a method of eatimating machinery and machinery related expenaea haa been queationed (Mott, 1985). However, while potential weakneaaea do axiat, USDA generated data ia the moat complete available and reaaonable concluaiona can be drawn from analyaia baaad on thia data. For the purpoaea of thia theaia, the USDA'a coat of production atatiatica will be uaad rather than any Telfarm data. The raaaon for thia hingea on three pointa. Firat, the atatiatica are gathered at the atate level in a conaia- tent faahion. Thia allowa for interregional and even interatate (when the original data ia available) compar- iaona to be made without aignificant error. Second, uaing USDA atatiatica enablea eaay compariaona to be made with other atudiea addreaaing aimilar iaauea. Finally, although variation ia praaant between the USDA and the Telfarm atatiatica, when definitiona are atandardized between the two aourcea, total coat of production for milk in Michigan comea within one percent of each other for 1983 data (Mott, 1985). 2.6.2 Coat of Production The key elementa of total coat of production are the fixed and variable expenaea. Variable coata will increaae 29 aa total production riaea, while fixed coata will not, gggggig_pgzigu1. Coata of producing milk vary widely over time from farm to farm and atate to atate due to differing production levela per cow, climatic conditiona, management practicea, herd aize, feed pricea and labor to name a few of the many influencea. In general, the Upper Midweat (Michigan, Minneaota, South Dakota, and Wiaconain) haa higher fixed coata than other ragiona becauae of heavy inveatment in buildinga and harveating equipment. 0n the other hand, the Upper Midweat enjoya a lower variable coat than other ragiona becauae ita dairy farmera produce a large portion of their own feed, utilize more family labor and have a higher average output per cow (due in part to higher quality forage). Feed coata repreaent a very important element in dif- fering regional coata. Both Appalacia (Kentucky, North Carolina, Tenneaaee, and Virginia) and the Southern Plaina (Texaa) are grain deficit ragiona and aa auch face large axpandituraa on imported feed. Additionally, the forage grown in theae two ragiona ia generally of a poorer quality relative to the Upper Midweat, again requiring producara to pay relatively higher feed coata per hundredweight of milk produced. For example, in 1984 the Upper Midweat'a feed coata accounted for 41 percent of the region’a total variable expenaea aa compared to 55 and 51 percent for the Southern Plaina and Appalacia reapectively (USDA, Sept. 1985). Figure 2.2 helpa to illuatrate thia point. 30 (Olcwt) 12 114 "31 N/ V 5 7i 6'1 5% ‘44 3'1 24 ‘1 0 u . All Appalacia Corn Belt M. Beat 8. Plaina Mid-weat (\m L W *W 1221 Feed {SI Other Variable -Fixed Figure 2.2. Coata of production: variable and fixed, by region for 1984 Source: m t rm o ' Emm- USDA. 1985 . 2.6.3 Michigan’a Relative Poaition In contraat to theae high feed coat ragiona, Michigan producara characteriatically produce the majority of their feed requirement in the form of haylage and corn ailage. Thia aelf aufficiency haa the effect of lowering feed coata and, with a alightly higher quality forage, helping to increaae production per cow. Together theae influencea provide Michigan farmera with a lower variable coat per hundredu Conperod Upper 111' cost: at the rela selected Table 2. Image“ Georgu KOHtUCky IlChlgan linnuot: I Carolin TOnnegge‘ Virginie “Moreen d°tl fer lichi. over 50w Illchl91m "it“, . ”innelot‘ slightly with ' t< 31 hundredweight than experienced by producara in the South. Compared with the two other primary producing atatea in the Upper Midweat, Minneaota and Wiaconain, Michigan'a variable coata are generally alightly higher. Table 2.2 illuatratea the relative coata of production and competitiveneaa of aelected atataa. Table 2.2. Coat of Production, Price, and Return to Riak Management for Selected Stataa, 1982 (m/cwt) Fixed Variable Total Price Return to State Coata Coata Coata of Milk Riak Mgt Georgia 3.46 9.09 12.55 14.40 2.42 Kentucky 3.35 8.14 11.49 13.50 1.32 Michigan 3.61 7.12 10.73 13.60 2.15 Minneaota 4.06 6.42 10.48 12.98 0.25 N Carolina 3.15 9.23 12.38 14.70 1.90 Tenneaaee 3.03 8.55 11.58 13.60 0.94 Virginia 3.29 8.86 12.14 13.90 1.29 Wiaconain 4.03 6.28 10.30 13.22 0.03 Source: USDA, "Firm Enterpriaa Data Syatem”. Unpubliahed data for 1982. Michigan producara do not hold the aame clear advantage over Southern atataa in fixed coata. However, even though Michigan’a fixed coata are higher relative to all Southern atatea, they are not aignificantly higher. Compared to Minneaota and Wiaconain, Michigan’a fixed coata are alightly lower. The lower variable coata provide Michigan with a total coat of production below that of all Southern atatea, ea demonatrated Figure in 2.3. (I/cvt) Figure : 1982. $014ch: “ta for Basec °°'Par1‘ nichigar 1n ulk total Cc the UPPe offlinne. relative ‘tetga (a) N “2:: "2% ‘&W ‘W s-\\\\\V/////////////fl *‘W § 4 .. 3. ,./ o . . /// 0A KN me me we 22 find Coata SS Von-ecu. Coat Figure 2.3. Total coat of production: aalactad atatea, 1982. Source: USDA, "Firm Enterpriaa Data Syatem". Unpubliahed data for 1982. Baaed on the data praaantad here and the aubaequent compariaona made, it can be aeid with confidence that Michigan holda a competitive advantage over Southern atatea in milk production. Futhermore, when the difference in total coat of production between Michigan and the rest of the Upper Midweat, apecifically the heavy aurplua atatee ofMinneaota and Wiaconain, ie coupled with Michigan's relative proximity to the high coat of production deficit atatea, it appeara Michigan may be in a poaition to uteblu e the w mcreau which em being qu. to benef; 2.7 Rec: The re fat he: h ereee um thch hlv u“ of co prim“), cultured . lilk. Th Iilk ind . Carolin. “-th of “I" dug Bum,u “1 There ‘ which hm M “d 1: of bulk TI \ 6 A “at” e P “d “pm, to I 33 eatabliah itaelf aa a key auppliar to Southern atataa. At a time when technological developmenta promiaa large increaaea in productivity and expenaive dairy policiea which encourage inefficienciea within the induatry are being queationed, Michigan producara could be in a poaition to benefit from their competitiveneaa and location. 2.7 Reconatituted Milk The reconatitution of nonfat dry milk powder and milk fat haa been done for many yeara throughout the world in areaa which are either great diatancea from auppliea or which have aupply conaiatency problema. In the U.S. the uae of concentrated milk, including dry milk powder, haa primarily been limited to the production of cheeae, cultured buttermilk and, to a amall extent, fluid beverage milk. Thia latter market ia aupplied by “raconatitutad" milk and haa been limited to a aingle proceaaor in North Carolina marketing a blended milk products, and to the atate of Alaaka, where approximately one third of the fluid milk demand ia aupplied by raconatitutad milk (Hammond, Buxton and Thraen, 1979). There exiat aeveral different mathoda of concentration which have been applied to fluid milk. With the advent of new and improved tachnologiaa the liat of potential mathoda of bulk reduction available haa expanded. Specifically, 5 A product made from a blend of freah whole milk, water and nonfat milk powder allowing the area’a fluid milk aupply to be atratched. eelhre Thee I currer of co: Iilk : are e eddre Ed eever inher dryer requ} lek varll lllk high (UP) eo.. 'dVQ trat he. chin °f t: 34 membrane aeparation techniquea have been gaining intaraat. Thia atudy deala directly with a product which would currently be claaaified aa raconatitutad. Aa auch, mathoda of concentration, uae reatrictiona impoaed on concentrated milk forma, and previoua reaearch addreaaing theae iaauea are all of importance. The remainder of thia chapter addreaaea theae areaa. 2.7.1 Methoda of Concentration Edward Jeaae (1980) preaented a atudy which liated aeveral mathoda of bulk reduction and aome of their inherent weakneaaea. The more traditional mathoda, apray drying, conventional evaporation and thermal evaporation, require large amounta of liquid fuel. Aa a raault, the milk undergoea a phaae change7 reaulting in aubatantial variation in the final product. A aomewhat leaa traditional approach to concentrating milk involvea the uae of reverae oamoaia filtration (R0) either on ita own or in conjunction with ultra filtration (UP). Currently, the two-atage UF-RO procaaa ia uaad commercially in the concentration of cheeae whey. An advantage of filtration over thermal-evaporation may be in ita reduced liquid fuel requirementa. Additionally, fil- tration ia leaa detrimental to the aolution'a conatituenta becauae in the abaence of heat, no phaae change occura. 7 A phaae change occura when a diacrete homogeneoua characteriatic of the aolution ia aeparated from the reat of the aolution by aome external force. To dot lethode o «cording Fro: thie nulberl : I1lee. II evaporet of conce drying t greater t° be at '“Perior high rel AdVQr '“hancec technlq‘ thm: ‘ “Mum “in.“ WM. 2.7 . R'Vel produCt "gulat' “mm 35 To determine the economica of employing theae varioua mathoda of concentration, they were ranked by Jeaae according to their proceaaing ahipping and aaaembly coat. From thia a tranaportation coat aurfaca waa developed. The numbera indicated that for diatancea greater than 100 milea, aome form of concentration ia economical. Thermal evaporation waa found to be the moat coat efficient method of concentration for diatancea up to 900 milea with apray drying to dry ingredienta becoming moat economical at greater diatancea. Membrane aeparation mathoda were found to be attractive due to their non-reliance on heat and auparior taate, however, their coat atructure waa rather high relative to the more traditional techniquea. Advancementa in membrane technology have greatly enhanced the potential for applying membrane filtration techniquea to the concentration of whole fluid milk. Thia theaia deala aolely with the uae of R0 filtration which ia diacuaaed in greater detail in the following chapter. The current regulatory environment faced by raconatitutad milk producta are examined below. 2.7.2 Uaa Reatrictiona Ravaraa oamoaia filtration of fluid milk producea a Product which, under current claaaificationa, would be r‘gulated aa a raconatitutad milk product. Aa they ‘3‘Errently atand, the regulationa for raconatitutad producta would Off edopting Under do e wore Peereurr: Service 1 product: recoehrn. (USPHS, ; product 1 Ilount o; Alloc: Pricing 1 tion of the ‘ark ”ovum f'd|rq1 Drovmc hem" or 'qUOJ 36 would effectively eliminate the economic incentive for adopting tachnologiaa auch aa R0 filtration. Under current government regulationa, “raconatitutad“ ia a word ahrouded by negative connotationa. The Grade “A“ Paateurized Milk Ordinance of the U.S. Public Health Service definea raconatitutad or recombined milk and milk producta aa ”...milk producta...which raault from the recombining of milk conatituenta with potable water" (USPHS, 1965). Given thia definition, any blended milk product muat be labeled "recombined" regardleaa of the amount of water which waa added to ”reconatitute" it. Allocation proviaiona, compenaatory paymenta, and pricing proviaiona impoae aignificant penaltiea on produc- tion of raconatitutad milk. Additionally, reatrictiona on the marketing of raconatitutad milk and mandatory pricing proviaiona for ingredienta axiat under both atate and federal marketing ordara. For example, the pricing proviaiona held by eleven atataa act to inaure that the handler producing raconatitutad milk will pay greater than or equal to the local Claaa I fluid milk price for the ingredienta going into raconatitutad milk (Hammond, Buxton and Thraen, 1979). Allocation proviaiona aaaign recon- atituted milk to the loweat uae claaa regardleaa of ita and uae. For every quantity down-allocated an equivalent amount muat be up-allocated. If there exiata inaufficient reaourcea to ”up-allocate“, a compenaatory payment equal to the Claaa I differential ia charged to the exceaa quantity . of recon pay the iron Clo which he have ex; below. Feder deergnec lenufect eecond ; would h. “term the pore Conlunpt tllred pg “Pact ¢ In 15 °f them “Dr. "Met. ““9- b. 37 of raconatitutad milk. In eaaence then, proceaaora muat pay the local Claaa I price for raconatitutad milk made from Claaa III or Grade B producta. Several atudiea, which have reaearched the impact of theae regulationa and have explored aome policy altarnativaa, are diacuaaed below. Federal milk marketing order price differentiala are daaignad to reflect tranaportation coata. Charging the manufactured price for the condenaed milk or developing a aecond price differential atructure for raconatitutad milk would help to eliminate thia gap. In a atudy of theae two altarnativaa it waa aatimatad that raconatitutad milk had the potential for capturing over one third of the fluid conaumption when ingredienta were priced at their manufac- tured price: the aecond alternative having a leaa aevere impact on farm pricea (Whipple, 1983). In 1979 Hammond, Buxton and Thraen publiahed the reaulta of their reaearch on the potential impact of raconatitutad milk. They inveatigated what effect reatrictiona placed on raconatitutad milk would have on regional price relation- ahipa between fluid and manufactured producta, am well me on production, uaage, and conaumer/producer welfare. Specifically, they looked at two altarnativaa to the preaent pricing acheme: 1) alter the differentiala while continuing to uae the aaaigned Claaa I price for fluid and raconatitutad milk: and 2) maintain regional differentiala between fluid and manufactured pricea by way of a Claaa I price on ueed in 1 Their reconetit price of price VOL atudy (tr coeperiec incoeee u would be Pruductic would inc '9le pm Sillle (1981) ., “ring “I and "Co: "Ui- r. commg m “n. fleet“ Prom 10 “Wed, than fr. ”Meg, 1' pro"lilo: ”(let v1! 38 price on freah fluid milk but remove auch pricing for milk uaed in ingredienta for reconatitution. Their reaearch indicated that under acenario one, raconatitutad milk producta would have little impact on the price of milk going to manufactured uae and the Claaa I price would decline in three of the eight ragiona in their atudy (the Hortheaat, Southeaat and Southcentral). By compariaon, under acenario two they found that producer incomea would decreaae more and the fall in Claaa I price would be greater, aa would be the change in utilization and production. Furthermore, the manufactured gooda’ price would increaae, producer incomea would decreaae and once again purchaaea by the CCC would be eliminated. Similar atudiea were carried out by Novakovic and Aplin (1981) and Novakovic (1982) who employed an economic engin- eering approach to determine the coat of producing blended and raconatitutad milk relative to the atandard freah milk. Their reaearch reaulta indicated that the coat advantage of ahipping and reconatituting/blending the milk waa aenaitive to the Claaa I price: the advantage being greater in markata with high Claaa I pricea. When FMMO proviaiona reatricting the aale of raconatitutad milk were removed, the coat of producing blended milk became leaa than freah milk proceaaing coata in many major markata. Hence, in the abaence of current reatrictive pricing proviaiona, coat incantivaa to reconatitution were found to axiat virtually everywhere. The aatimatad retail level out of t eilk rang oente in 1981 pric provieior otherwiae Recent (prieeril ‘P‘cifice «v.11 . 00 the pi Proceee, main M “1“ Deli hm pub “he. Ch‘pt‘r M E The "“3 c: end M “at“ new." Emil!!!) Richly” l no”We“ 39 coat of theae regulationa reatricting uae of raconatitutad milk ranged from 3.2 centa in Chicago to a high of 10.8 centa in Boaton for the ahelf price of a gallon of milk (in 1981 pricea). Clearly than, the current FMMO pricing proviaiona aignificantly eliminate incantivaa which would otherwiae axiat in favor of raconatitutad milk. Recently there have been aeveral economic atudiea (primarily out of Cornell Univeraity) which have looked apecifically at the application of UF at central locationa aa well aa at the farm level. With the rapid improvementa on the pivotal componenta of the membrane aeparation procaaa, intaraat in theae bulk reduction altarnativaa will remain high. In addition to theae atudiea on raconatitutad milk policy iaauea, aeveral apatial modeling atudiea have been publiahed which approach theae and other policy areaa. A diacuaaion of auch atudiea ia preaented in Chapter Four. 2.8 Summary The natural characteriatica of milk have led to a unique and complex aet of inatitutiona daaignad to provide a aafe and aecure aupply of milk in an orderly faahion. At the center of the FMMO ayatam ia the exiatence of the Upper Midweat’a competitive advantage in the production of milk. Examining the production coat atructure indicated that Michigan itaelf holda a competitive advantage over the non-Midweat atataa with which it would potentially compete. Exeeini within the colpetitii first ete' iiltretic eppeere I State di mcmu “reconhi tion, h Further recoebi accept; feeaibi technol the Mr] Four on: M filtr {Olloun 40 Examining the relevant market inatitutiona exiating within the dairy induatry and eatabliahing Michigan’a competitive poaition in milk production were fundamental firat atepa in addreaaing the potential impact of R0 filtration of fluid milk. At preaent, reconatitution appeara economically viable in the abaence of Federal and State diaincentive regulationa. If R0 filtration ia uaed to concentrate fluid milk, with the concentrate then being "recombined" nearer the point of final aale, tranaporta- tion, handling, and atorage coata will be reduced. Furthermore, R0 filtration providea the moat likely recombined or raconatitutad fluid product for conaumer acceptance. However, before auch generalized claima of feaaibility are made, a more detailed deacription of the technology ia neceaaary and the effecta of R0 filtration on the marketing of fluid milk ahould be conaidered. Chaptera Four and Five addreaa the latter of theae iaauea while the R0 filtration procaaa and technology are praaantad in the following chapter. CHAPTER 3 REVERSE OSMOSIS FILTRATION Reverae oamoaia (R0) filtration technology waa developed in the 1950a. Initial reaearch focuaed on uaing R0 to produce pure water from aea and brackiah watera. In more recent yeara R0 haa been uaad to filter apple concentratea, uaad in candy manufacturing and, within the dairy induatry, it haa been uaed to procaaa whey. Preaently within the dairy induatry, RO filtration ia aeen aa a reaaonable alternative to evaporation, requiring leaa heat and not involving a phaae change. Additionally, RO filtration haa the potential for becoming an on farm procaaa. Although under current conditiona on farm RO ia not economical, given greater apecialization, improvementa in the tech- nology, and increaaed hauling chargea, on-farm RO could find future uae. Thia chapter preaenta the technical aapecta of employing RO filtration for the concentration of fluid milk. Throughout all diacuaaion of RO filtration within thia atudy, it ia aaaumed that the whole fluid milk will paaa through a aeparator prior to filtration with the cream than 41 reeixed he eppli 3.1 Pr. The l regerdl. tekee p Ieperet. through IOlutea 0510' between 1' Peret tolvent the 301 lllu‘tr that Hh DIOCQ‘E achim leebren illuItr Given ' 'e‘bren “hi“. (011“le 42 remixed before ahipping. Hence, the actual filtration will be applied to akim milk. 3.1 Principlaa of Reverae Oamoaia Filtration The baaic principlea behind RO filtration are the aame regardleaa of the aolution involved. The oamotic procaaa takea place in all organiama where water and a aolution are aeparated by a membrane and the water naturally paaaea through the membrane to dilute the aolution. Hence, all aolutea in aolution exert an oamotic praaaura. Oamotic praaaura ia needed to maintain equilibrium between the aolution and the water acroaa a membrane, which ia permeable to the aolvent (i.e. water). The flow of the aolvent from ita pure atate through the membrane and into the aolution (i.e. milk) ia termed oamoaia. Figure 3.1 illuatratea thia principle. The maximum work involved ia that which ia againat the oamotic praaaura. The oamoaia procaaa will continue until a atate of equilibrium ia achieved: hence, the preaaurea exerted on both aidea of the membrane are equal. An example of thia procaaa can be illuatrated with aea water, which ia 3.5 percent aalt. Given a atrong membrane, the movement of water through the membrane would continue until a column 750 feet high ia achieved: equaling the oamotic praaaura of the aalt (Dunkley, 1971). Figure Equi 0510.1. th! oer tional Oh hYd] °|l031p “gut. 43 SOLUTION WILL RISE TO THIS - — — - 4— POINT WHICH IS HEAD EOUALTO APPARENT OSMOTIC PRESSURE SEMI-PERMEAB LE MEMBRANE MN MORE LESS CONCENTRATEO CONCENTRATED SOLUTION \ SOLUTION ‘\ WATE R F LOW Figure 3.1. The oamoaia procaaa Equilibrium impliea reveraibility. To reverae the oamoaia procaaa, aufficient preaaure muat be applied until the oamotic praaaura of the aolution ia overcome. Addi- tional preaaure muat then be applied in order to create. an hydraulic flow in the oppoaite direction. The reverae oamoaia procaaa ia illuatrated in Figure 3.2. Pnessuae AAIA semesnmuaie MEM;RANE none LESS CONCENTRATED CONCENTRATED SOLUTION \ SOLUTION \X 7 7 ' \ WATER now Figure 3.2 The reverae oamoaia procaaa 44 R0 filtration can be thought of aa a praaaura driven membrane procaaa in which aubatancea are fractionated, aeparated or concentrated without the aubatance undergoing a phaae change. The minimum work for thia procaaa ia the exiating oamotic praaaura. Hence, in R0 filtration the driving force ia the net raault of the reaiatance and oamotic preaaure on both aidea of the membrane and the applied praaaura. The relevant oamotic praaaura involved ia not that which 1a in the bulk of the aolution, rather, it ia the oamotic praaaura at the aurfaca of the membrane. In the filtration of milk, thia becomea increaaingly important over time. Aa water paaaea through the membrane, a layer of continually incraaaing concentration developa on the membrane aurfaca. Thia layer, primarily caaein, reducea the effectiveneaa of the membrane1. 3.2 The Reverae Oamoaia Syatem RO membranea muat be mounted in equipment providing the neceaaary aupport and flow control. There are four main variationa in deaign with the general principle remaining the aame. The meat common ayatam involvea mounting tubular membranea in a aeriea of poroua aupport tubea connected by headera. A croaa aection of one of theae tubea ia illua- trated in Figure 3.3 and the general R0 ayatam praaantad in Figure 3.4. 1 Caaain ia the protein component in milk. PERVE FLOW {i Figure oeeoe: Source Fisu; 45 SPIRAL MODULE p ERMEATE ADHESIVE BOND CONCENTRAT E .fl'PéN—r' FEED FLOW TRATE EAIE CONCEN PERM now PERMEATE CARRIER MESH SPACER MATE RI AL GRING MEMBRANE Figure 3.3. Standard membrane module uaed for reverae oamoaia Source: Oamonica, Inc. Minnetonka, MN. Hydraulic Accuaulator Pres 6a sue me BumIkunme Q Regulator vi". 3: menu! hmp Human“ InnueOnnns nmenmue Concentrated Solution rum Solution Hater hummnm! Hatar r Control Inflow - “/ Tum Figure 3.4. The reverae oamoaia ayatam 46 The aolution (i.e. milk) to be concentrated paaaea at high velocity through the tubea where the filtration takea place. The reaulting permeate then travela through the membrane before it ia collected outaide the ayatam of tubea. A primary requirement for filtration ia for the aolution to be circulated at very high velocitiea within the tubea to prevent the concentrated layera from forming on the membrane aurfaca, fouling the ayatam. To achieve the required aolution velocity, a praaaura pump capable of producing up to 2000/pai throughout the ayatam ia required. Additionally, turbulence promotara or volume diaplacing roda are commonly uaad to generate increaaed turbulence near the membrane aurfaca. 3.3 Key Factora Influencing Operation of the RO Syatem. The prevention of the caaein build-up on the membrane aurfaca haa been one of the maJor areaa for technical improvement in recent yeara. Both the membranea and the cleaning mathoda have been improved, aignificantly en- hancing the proceaa’a economic feaaibility. Thua, the membrane and the flow through the membrane, termed the flux, along with aolution velocity, praaaura, concentration level and temperature, are all important elementa of the ayatem’a operation and merit further diacuaaion. Figure 3.5 providea an over view of the key factora influencing the efficient operation of the R0 filtration ayatem. 47 .2: .2333: “~53 ..1 a» suzncoucom “:3 .53ch "nouuaom Eon» «queue down? 3:080: 080. no. 83:. comuauuzu On «A» «a camuauomo adduced—«om :ouuau mou— .n.n enema”.— Ron can» aouaouu aucaunBuE «o x:—« nun-eunuc- aomuauucuucou acmcao—u «convoumu wn~u aomuaqou aomuuuma-uoo an oeaumamnoud um u-epu— Seamueo om uuaouuuv .o-auuucnu ~o>oa one upmacoowo o» «uae-nudes» RN a o. apnea nun-.uuo com-nu» am euoeaouqocou no .ouonueue comuauucouoou ..coMSSSOn nooucoU anon» udmucoao—oau .sumuodoe unmade cw o-aouuom Ru < nouaouucuu .uoz comaa~om auuou $~ a» a ma x=_u uuuuuuuamn udfiuaueeo aeaaouuum .m neeueov oo ma~u ooaouuam comuauuuuuoOu aaaoauumacou on oomuouo. uqmi .980» a“ comunqo. menace cue-uauan an» mama-0:: coma cannon n < nouaouuuae ..om oucuauuoaoh memoooN .uoumee «mau— eon-uaaua aaa0u sound» «a Boa.»- o» a: uae-none aeaaouuom uae.-one omaaomaa on new) uncouucm new: nonmacou nah ease-uae «a 0.3: —~m3 «aaa wa_m wo_u nouaouuomu ..om «Suwanee noon women—e ~u>ou comuanmuauoe name uae—o» unannouau a camuauucoucou eueou noun-«one no euuuofioue o» no humuo_u> nouauuuuVI comma aoaaouucm coca—onus» new? uuaouucm comuauauoe we acoaemaco uncoqumvo< no aomua—_auuc~e __m3 ouau w=_m oaau non-ouucuo ..om zuuuo~u> .55: m5. 2.2 3 3.: mZO—haudmxu Ungm mggm~8m¢ Pummmm m0 mF' ‘$‘ _. 5.- 3- / ‘E’ I« — e e e :~ 0 5 IO I5 20 . PROTEIN IN CONCENTRATION (3) Figure 3.9. Effect of temperature and percent protein on relative viacoaity of the concentrate Source: Goldamith, et al., "Recovery of Cheeae Whey Proteina Through Ultrafiltration." Waahington D.C., November 1970, p.4. An advantage for on-farm RO concentration may axiat due to the optimal temperature for R0 filtration of milk under 55 aome conditiona being approximately the aame aa the temperature of milk coming directly from the cow, 90 degreea fahrenheit (Jeaae, 1980). Thia fact may be important if the location of the R0 filtration facilities ia queationed: on-farm or in a central proceaaing plant. 3.4 Summary The dairy induatry ia going through a period of major adJuatment. Aa apecialization continuea within agriculture and the average dairy operation increaaea in acale, produc- tion may become more regionalized. Thia will require economical tranafar of a whole milk product for fluid conaumption which meeta all aapecta of conaumar damand. RO filtrated milk appeara to aatiafy theae requirementa. Both the R0 ayatam deaign and ita membrane are contin- ually being improved to enaure greater operating effic- iency. The noncelluloaic membranea available today, being more durable, allow more coat effective cleaning mathoda to be employed and yield a longer life. Thia helpa reduce the overall coat of operation and promotea the RO filtration proceaa to an appealing level. With increaaed uae of RO filtration in the dairy induatry, further improvementa would likely be made. Thia, together with conaumar accep- tance, indicatea that RO filtration of milk for fluid conaumption could have a futrue role within the dairy induatry. CHAPTER FOUR THE MODEL Thia chapter addreaaea the application of apatial equilibrium modeling to the dairy induatry. A brief diacuaaion of the advance of apatial equilibrium modeling of agricultural commoditiea ia given followed by a daacrip- tion of the apecific apatial equilibrium program and aolution algorithm uaad within thia atudy. Additionally, the general modal’a equilibrium conditiona, price linkage function and limitationa are aet forth. A review of previoua economic reaearch in the dairy induatry ia included and finally, the actual model and data incor- porated into thia atudy are praaantad. 4.1 Spatial Equilibrium Modela Spatial equilibrium modeling providea the appropriate avenue for analyzing the impact of aalactad changea within the dairy induatry. Thia form of modeling, which endog- enize trade flowa and market aharea, haa been uaed exten- aively for comparative atatica analyaia of exogenoua variablea, auch aa policy changea. Of apecific importance to thia theaia, thia form of modeling ia alao efficient at 56 57 determining the effect on the net poaitiona of trading ragiona due to changea in tranaportation coata. 4.1.1 Review of Selected Literature To obtain apatial equilibrium in the general caae, pricea muat be found which will produce equilibrium quantitiea and price differentiala in and acroaa all markata deaignated within the model. Samuelaon (1952) firat approached thia problem by maximizing the aum of the areaa under the exceaa demand curvea for the importing regiona, leaa the area under the excess aupply curve for the exporting regiona, leaa tranaportation coata. Thia formulation, deaigned for objective function maximization, provided a apatially competitive equilibrium aolution. Takayama and Judge (1964 and 1971) were key contributors to this area, applying atandard quadratic programming methoda to Samuelaon’a model. Later thia model was enhanced to aolve acroaa perioda (Takayama and Liu, 1975) and to aolve optimally for multi-commodity trade (Takayama and Haahimoto, 1976). Among commoditiea atudied, feed graina have received the moat application of the one-commodity, one-period model. Beginning in the mid-aixtiea, empirical research developed which applied quadratic programming, apatial-price equilib- rium modela to world agricultural trade (Schmitz, 1968: Chang, 1972: and McGarry, 1968). Linear programming, 58 quadratic programming, and network flow models became standard tools for analysis. These traditional model formulationa have most commonly been used to ascertain optimal freight flowa. However, while they have been used extensively, they are subject to several limitations. Primarily, linear programming and network flow models are constrained to having linear obJec- tive functions, while quadratic programming models require linear export supply and import demand functions. More recently, researchers have begun to develop and apply non-linear spatial equilibrium models to trade in agricultural commodities (Warner, 1979: Holland and Pratt, 1980: Holland and Sharplea, 1984; Randolf, 1986). Not constrained by the limitation of linearity, these models have generated a great deal of attention. The spatial equilibrium algorithm used in this thesis falls within this class: the Huhn-MacHinnon Vector Sandwich Method algorithm (discussed in section 4.2.1). 4.1.2 General Spatial Equilibrium Model Each region in the general trade model represents a market with its unique supply and demand characteristics. In any given period, regions may be in a surplus or deficit standing with respect to the given commodity. It is the existence of surplus and deficit markets which creates trade across regions. The region: neceee: the be. price, titiee e decr coeeod coeeod the 1. in the 'XPect the d. Coeeod Cdjugt The ledel ‘ge Que Jth ‘Se Flu th. EQu net .1“ ‘Qu 'Qu “II“ R p. EQUJ Illu‘tz 59 The cost of transportation is central to the inter- regional trade problem. Any change in transportation cost necessarily will create a shock to the system (away from the base conditions) resulting in a change in commodity price, trade flows, and regional supply and demand quan- tities. In a free, open market, theory would suggest that a decrease in the transportation cost per unit of a given commodity would precipitate a price increase of that commodity in the exporting region and a price decrease in the importing region. The overall result being an increase in the quantity traded. In a regulated market one would expect the degree of adjustment to vary directly with both the degree of restrictions imposed on the free flow of the commodity and with the ability of the price mechanisms to adjust price. The general interregional spatial-price equilibrium model can be formulated as follows: The 1th source (exporter) is a collection of agents in region 1 who are willing to supply a quantity 01 at a per unit price of P1. In turn, the 3th sink (importer) is a collection of economic agents in the 3th region who are willing to pay P1 plus a transportation cost of T1) for the quantity they demand at price P3 (where P3 8 P; + T13). Equilibrium between each i region (those which are a net source) and each J region (those which are a net sink) exists when the regional price differential equals the transportation cost, T13. At that point, equilibrium quantity equals 0' and equilibrium price equals P“; in region i and P'J in region 3 where P“) I P“; + T1). Equilibrium across two apatially separated markets is illustrated in Figure 4.1. This figure represents the 'IQ‘t 60 single product, two region case in the form of a back-to- back diagram. Region Y’s supply and demand curves are plotted on the right side and region X’s are on the left side in reversed form. Each region’s excess supply curve is derived from regional supply and demand schedules, representing the amount by which the quantity offered exceeds (or falls short of) the quantity demanded. Deficit Surplus Market (X) Market (Y) Figure 4.1 Impact of transfer costs, 00’, on prices and trade between two apatially separated markets2 Source: Raymond G. Breasler Jr. and Richard A. King, W- 1978. p-91- 2 Where ESx I excess supply in region X: ESy I excess supply in region Y: S; I supply in region X: Sy I supply in region Y: Dx I demand in region X; Dy a demand in region Y: Ox I quantity supplied/demanded in region X; 0y I quantity supplied/demanded in region Y: P I price: t I per unit transportation cost. Equtl eupply c eupply c price ii cost, 0( eerket, equilibi 0f good that ill 4.2 Be: Hol; thou. 9p r'glone. progra. Iolvg. i Min .19 '°1Utio bQlou, Th. ‘190r1t H°°Kinn tarY‘pi. vsa ,0“ 'p‘tiel 61 Equilibrium price occurs at the level where the excess supply curve for market X, ESx, intersects the excess supply curve for market Y, ESy. Note that the equilibrium price in the two markets differs exactly by the per unit cost, 00’, of shipping the good from the excess supply market, Y, to the excess demand market, X. Furthermore, equilibrium at each price level ensures that the quantity of good 0 exported from region Y, f’g’, is exactly equal to that imported by region X, e'd'. 4.2 Generalized Transportation Problem Holland and Sharplea (1984: Holland, 1985) put the above general formulation into a form useable for inter- regional trade analysis. They developed a micro computer program, Generalized Tranaportation Problem (GTP), which solves the spatial equilibrium problem via the Kuhn-Mackin- non algorithm in the international trade setting. The solution algorithm utilized in GTP is briefly discussed below, followed by a more comprehensive look at GTP itself. 4.2.1 Vector Sandwich Method Solution Algorithm The Vector Sandwich Method (VSM) spatial equilibrium algorithm, developed by Kuhn and MacKinnon (1975) and MacKinnon (1975), solves via a fixed-point, or complemen- tary-pivoting algorithm approach. A primary benefit of the VSH formulation over the traditional method of solving spatial equilibrium problems is that VSM is not bound by 62 the restriction of linear excess schedules. It can handle non-linear demand and supply relationships and can even accommodate non-smooth (first derivative discontinuous) functions. VSM solves by searching directly for equilibrium prices and quantities which will satisfy a specified set of equi- librium conditions. The process involves dividing the total solution space into a set of several aimpliciea. Then, using a sophisticated search procedure, it generates a ”path" which leads to an equilibrium point. In this way VSM yields the equilibrium conditions for the general interregional trade model. 4.2.2 GTP Equilibrium Conditiona GTP is not capable of producing an exact solution, rather, it generates a solution to a piecewise linear approximation of the system of equations defining the original problem. While it is true that the solution generated is not exact, estimated to eight decimal places, it is satisfactorily accurate. GTP’s solution procedure is based on a set of equilibrium conditions which are subject to a set of constraints -- some innate to the program and others which may be manually specified by the operator. The model’s general equilibrium conditions (items 1, 2 and 3) and constraints (items 4, 5, and 6) can be stated as follows: 63 (1) For each region the amount supplied of the commodity is defined by the quantity dependent excess supply schedule and must equal total out-shipments. (2) For each region the amount demanded of the commodity is defined by quantity dependent excess demand schedule and must equal the total in-shipments. (3) The amount in total which is supplied across all regions must equal the total demand. (4) Supply and demand schedules define convex, non-empty feasible solution sets in non-negative price-quantity space. (5) Assumption of free disposal applies, ensuring that the equilibrium price will be non-negative. (6) Total excess demand will become negative if the sum of prices is sufficiently large. Conditions (4) and (5) are implicitly in the model while condition (6) is explicitly in the model as part of VSM. 4.2.3 Price Linkage Mechanism Within GTP there exists the ability to account for regional price differences through the specification of a price linkage function. This function acts to incorporate the relationship between exporter price and importer prica via a combination of potential price wedges such as tariffs (ad valorem and specific), exchange rates and transpor- tation costs. Hence, when no tariffs exist and the same currency is used in each region, in equilibrium the differ- ence in price of the traded good between regions is exactly equal to the cost of transportation. The price linkage function can be specified as follows for trade between the ith exporter and the 3th importer: 64 L3(P1) - (((P1I(V1 + 1) + U1)IE1 + 713>- 100 (4.3b) H < 100 (4.3a) 8.30/cwt 100 miles M > 100 (4.3c) H < 100 (4.3a) 100 > M > 378 (4.3b) 8.90/cwt 378 miles M > 378 (4.3d) H < 100 (4.3a) 100 > M > 772 (4.3b) 81.75/cwt 772 miles M > 772 (4.3a) The break even distance function for increasing costs of RO filtration is presented in Figure 4.3. The function indicates that for distances greater than 100 miles, application of RO filtration is profitable. Analogously, for any cost of RO less than 8.30/cwt, RO filtration is profitable at any long haul mileage. NO 7004 and mod good 300+ 200- One-Way Hiles 100 of‘ I Y V U W T T Y T 0.3 0.5 0.7 0.9 1.1 1 .3 1 .5 1.7 costs of RO filtration Figure 4.3. Break even mileage function under normal costs 79 4.3.6 Data Requirement The data required to run the model is dependent upon how the various parameters are defined. Efficiency, ease of operation, consistency with previous studies, and . coincidence with research goals were the primary criteria used when defining model parameters. Each parameter is discussed below with its data requirement and sources outlined. 4.3.6a Regions GTP is limited to ten exporting (source) and twenty-five importing (sink) regions. For the purposes of this study, each region is included as both a source and a sink, except for the tenth sink region which has been reserved as the non-Class I sink. Hence, a maximum of nine Class I regions is possible. Regions for this study were defined along state borders. The 3oining together of states to create regions was based on regional importance in the study, each state's 3uxtapoaition to its maJor export destination(s) and/or import aource(s), compatibility with previous studies, and data considerations. In all, 29 federal orders and 33 states are incorporated into nine regions. The primary data source used in defining these regions was the FMMO, State of Origin statistics for Class I milk pooled under federal orders (USDA, FHOHS, March, 1986). A map of the regional boundaries is presented in Figure 4.4. .I‘ a Figure 4.4. Delineation of regions used in this study 4.3.6b The Tenth Region While Supply data is available for Grade A milk, only a portion of Grade A milk goes to Class I use. If the model were run with this Grade A data, yet only Class I demand, the excess supply of Grade A over Class I demand would «swamp the market forcing the equilibrium price down. In this sense, Class III demand must be incorporated into the -Odel within a balancing role. The tenth region serves tJ'i:l.s role. Specifically, it is given a near infinite d-land elasticity for Grade A milk at the M-W price. In this way, Grade A milk in excess of fluid demand is -1 located to Class III use. 81 In contrast to the other regions, region ten does not have a specific location: rather, it is assumed that every region has a Class III market which is located at each region’s supply center. Hence, the transportation cost from each demand center to region tan is set at zero. Such a specification further ensures that milk bound for Class III use will neither be imported nor condensed through reverse osmosis. 4.3.6c Regional Centers A market supply and demand center has been designated for each region. This center serves as the base point for demand/supply, price, and shipment costs between regions. Regional market demand centers were designated as the closest ma3or city to the estimated center of population for the region. Population data for 1980 was used to estimate the market demand centers (USDC, 1983). Regional supply centers were determined by looking at a milk cow numbers map and selecting the closest city to the estimated center of milk cow population. Data and maps eon milk cow numbers were taken from the 1982 Census of Agriculture (USDC, 1985). Regional supply and demand c-nters are listed in Figure 4.5. 82 Region States Demand Supply Encompaased Center Center 2 Alabama Macon, GA Newnan, GA Georgia Mississippi South Carolina 3 Arkansas Lufkin, TX Greenville,TX Louisiana Oklahoma Texas 4 Illinois Galesberg, IL Ottumwa, IA Iowa Missouri 5 Kentucky Mount Airy, NC Hazard, KY North Carolina Tennessee Virginia 6 Indiana Newark, OH Mansfield, OH Ohio - West Virginia 7 Connecticut Port Jarvis, NY Oneonta, NY Delaware Maine Maryland Massachusetts New Hampshire New Jersey New York Pennsylvania Rhode Island Vermont 8 Minnesota Eau Claire, WI Eau Claire, WI Wisconsin ‘9 Michigan Highland, MI Lansing, MI 10 All of above All of above ------ ‘ F 19ure 4.5. States encompassed within regions, regional c'Omand and regional supply centers 83 4.3.6d Length of Run Each variation of run in this study is based on a one period, annual model. In dairying this is effectively a short-run period: neither producers nor consumers would fully ad3ust to a change in price during that time span. Furthermore, this length of time yields results based on the average market conditions throughout the year. Such an average must be considered for the implementation of reverse osmosis filtration technology. Additionally, market data is conveniently available in annual form. 4.3.6a Base Period The base period selected is 1985. This represents the most recent year for which all necessary data is available and provides base data which reflects the current macro economic conditions as closely as possible. Two maJor influences on 1985’s data should be mentioned. First the dairy diversion program, initiated in September of 1984 and extending through February of 1985, effected the prices in some markets. Second, the St.Louis-Ozarks order was terminated in April, 1985. The termination of this order impacted the data only slightly. The data was ad3usted where possible to reflect these changes and the regional- ization and one year length of run should effectively mask their impact. Also masked are other year to year varia- tions in base-excess plans, direct delivery differentials or other similar ad3ustments within FMMOs. 84 4.3.6f Bounds GTP allows the operator to set upper and lower trade bounds for each region. This restriction helps to mimic actual market conditions and ensure realistic results. Given the low price elasticity of supply and demand in dairying, little variation in supply or demand is likely over one period. No regional trade quantity bounds are set for Class I use. However, it is assumed that in each market a minimum amount of supply goes towards Class III use. For each region this lower trade flow bound is set at ten percent of 1985 supply levels. This forced allocation to the Class III sink is designed to reflect the combined effect of local production of Class II products and the natural loss of Class I quality milk which occurs during marketing. Thus, this trade level restriction serves the purpose of ensuring a more realistic equilibrium price. 4.3.6g Prices The actual 1985 price for each region was calculated as the sum of the average base Class I price, regional Class I differential, and regional over-order premium. The calcu- lated regional prices and regional elasticities were used to generate supply and demand schedules. Additionally, the Class I fluid differentials were directly incorporated into the price linkage function. Class I price differentials remained unchanged for each region throughout the base year. 85 Class I differentials were not necessarily uniform across all marketing orders within a region. In such cases the region’s differential was estimated by weighting the separate differentiala according to the amount of milk pooled in each order. Price differential data and data on milk pooled under FMMOs was taken from k Or W «USDA. .... 1986: August 1986» Regional over-order premiums were generated in an analogous manner from data on over-order premiums for selected cities from Qgigy_fl§£kgt_§;g§igtigg, (USDA, Maarch 1986). The error involved in estimating regional prices in this fashion is believed to have an insignificant impact on model results. 4.3.6h Demand and Supply Schedules Rather than specifying each region as being either in a state of excess supply or excess demand, separate supply and demand functions are used for each region. This has the advantage of directly setting the supply and demand functions from the available data and reducing the restric- tions on exceaa schedule form. Furthermore, predeter- mination of a region’s final trade status is not required. For this study supply and demand schedules are entered under the constant elasticity format. By using this form, the schedules are easily derived from available data for each region. The elasticities employed were obtained 86 from other studies. Specifically, three main sources of short-run supply elasticities were available. Hallberg, Hahn, Stammer, Elterich and Fife (1978) derived short-run supply elasticities from a report by Hallberg and Fallert (1976) for nine regions nationally. For the latter study, regional short-run farm level demand elasticities were derived primarily from retail level studies by Boehm (1976) and George and King (1971)5. Later Hammond, Buxton and Thraen (1979) used short-run and long-run supply elasticities as generated by Hammond (1974) and long-run demand elasticities as developed by Fallert and Buxton (1978). Supply and demand elasticities available for fluid milk generally have not been exacting. It is not unusual for different supply schedule functional forms to be used for different regions, as well as inconsistent data sources, time periods, or method of ad3ustment across variables. Recently Huy (1986), has attempted to eliminate some of this variability by using a duality approach to estimating short-run supply elasticities7. This methodology avoids the characteristic over estimation found with elasticities 6 The farm level elasticities were derived by scaling up the retail level elasticities by an assumed elasticity of price transmission of .5. Reference to this proceedure is found in George and King (1971). 7 Specifically, the method Huy employes is a profit function approach to duality theory using Zellner's similarly unrelated regression. This entails developing one function from which the other functions are derived; regionalization is achieved through use of dummy variables. 87 generated by linear programming (Cilley, 1985). Addition- ally, Huy's elasticities cover twenty-nine states on an annual basis from 1981 to 1985 allowing for more freedom in region specification. For the purposes of this study, short-run demand and supply elasticities will be taken from Hallberg at al. and Huy respectively. While the long-run demand elasticities estimated by Hallberg et al. are all less than those obtained by Hammond, proportionately they are the same between regionsa. These elasticities were generated for the larger USDA regions, but are not expected to vary significantly within those regions. Hence, they can be transferred to the regions used within this study with acceptable confidence. Huy’s initial short-run elasticities are used because of their current nature, more exacting estimation procedure, and their availability on a state and regional level. It should be noted that proportionately Huy’s and Hallberg’a elasticities appear to be very similar. Furthermore, comparison of model results generated using the two sets of elasticities indicated that the model is relatively insensitive to a switch between the two. This is discussed in section 5.7.2. 3 Hallberg et al. suggest long-run demand elasticities will be 1.5 times greater than their estimated short-run demand elasticities. Hammond’s elasticities are uniformly approximately 2.16 times greater than Hallberg’s short-run elasticities. To fully develop regional supply and demand schedules, the function’s slope, B, needed. and intercept, a, terms are also Each region's a and 8 terms are generated from the price and quantity values existing during the base year for each state within the defined regions (USDA, 1986 pp. 40-43) a FMOMS, May The a and 8 values associated with the regional supply and demand schedules are listed in Table 4.2. Table 4.2 Regional Supply and Demand Schedule Slope and Intercept Values Demand Intercept Slope OIDOIJIDUAOJNH 3.: Supply Intercept Slopa 3.820 .959 5.431 1.423 -7.196 4.829 29.834 2.747 10.886 5.841 27.531 2.695 78.003 11.207 173.040 5.374 37.890 .999 25.710 37.994 63.867 54.396 48.330 49.049 127.668 24.554 24.990 42499.000 -e130 -3572.00 The required data was incorporated into the model presented in this chapter. This fully specified model provides the tool of analysis through which study ob3ec- tives can be achieved. results and analysis of them. Chapter Five presents the model CHAPTER FIVE ANALYSIS OF RESULTS The model specified in the previous chapter was used to analyze the Class I market impact of RO filtration under several pricing and policy scenarios. Each of the ma30r scenarios, and the results generated from them, are discussed in this chapter. Figure 5.1 provides a reference to the various model runs and Figure 5.2 presents a diagram of their incorporation. The primary questions asked under each alternative case are how supply and demand quantities are affected, how the distributional pattern is altered, and what is the resultant impact on costs and revenues. Results are discussed primarily in terms of all regions as a whole with five regions, 1. 2, 7, 8, and 9, being isolated for cross analysis. Any other regions signif- icantly impacted by a particular parameter change will be mentioned as warranted. Model generated prices, quantities and trade flows for all regions under all runs are presen- ted in Appendix C. To prevent any confusion in terms of scale, comparisons are made on a percentage change basis. Any change of less than one percent is considered insignif- icant in terms of the model’s sensitivity to minor changes. 89 90 8R03 BROS 8R0175 SEPT SR03 SR09 SR0175 CCC occaoe ND D86 D86RO9 TC2 T02R09 Description Initial annual run serving as the standard for comparisons. 1985 market conditions with no R0 applied. BASE specification with R0 incorporated at a cost of 8.30/cwt. BASE specification with R0 incorporated at a cost of 8.90/cwt. BASE specification with R0 incorporated at a cost of 81.75/cwt. Model run generated based on September 1985 market conditions. Serves as a base upon which RO feasibility during month when shipments are high. No RO applied. SEPT specification with R0 incorporated at a cost of 8.30/cwt. SEPT specification with R0 incorporated at a cost of 8.90/cwt. SEPT specification with R0 incorporated at a cost of 81.75/cwt. BASE model ad3usted for reduced CCC purchases. An import quote was placed on Class III milk, reducing purchased by 444 million pounds. CCC specification with R0 incorporated at a cost of 0.90/cwt. BASE model ad3usted for the full removal of Class I differentials. No RO applied. BASE model with 1986 Class I differentials substituted for the 1985 levels. No RO applied. D86 specification with R0 incorporated at a cost of 8.90/cwt. BASE model with a fifty percent increase in transportation costs incorporated. No R0 applied. TC2 specification with R0 incorporated at a cost of 8.90/cwt. Figure 5.1. Reference of model scenario titles and descriptions ranaportation oat Functio It. 91 Iefine ‘egiona, Regions upply and Deman- enter I 1 Input Supply Input Demand Elaaticities Elaaticitiaa Prices and Prices and uantitiea usntitiea All 13 G nera Tranaportation Coat Matrix. Generate Regional Iver-Order Premiums All ‘J September Market Conditions Reduce CCC Purchase a Alternative Supply Elaaticitiaa Increase Transportation Costs Gate 2 Introduce RO Filtration [Change Class I Differentials to Remove Class I Differentials Figure 5.2. Calculate Price Linkage Between Goto 1 Flow diagram scenarios into the model Supply and Demand Schedules nalyze and Cross Compare Results. Incorporate Scenario Changes Where Appropriate and Rerun 3 of the incorporation of various 92 Although great effort was taken to reproduce actual industry characteristics, many aspects of the dynamic, Grade A milk market could not adequately be captured within this model. A brief reminder of some of the inherent limitations associated with this model will be discussed before analysis of the results is made. 5.1. Caveats and Limitations Several areas of caution are inherent to both the model specified and the spatial equilibrium program upon which it is run. First, the model is specified as a short-run model. It produces rather sudden shifts to parameter changes, neither accounting for the industry’s ability to redesign policies nor for the long-run market response. The short-run solution provides what may perhaps be an extreme response in the absence of dynamic interaction over time. However, if RO filtration is economically feasible, a short-run model should indicate so. Second, the computer program generates a perfectly competitive solution for an admittedly “imperfect“ market environment. One would expect then that the solutions generated may deviate from those actually produced by a complex and dynamic market. There is no way to determine where along the spectrum of economic markets the Class I market would eventually find equilibrium under the par- ameter changea discussed within this chapter. Hence, within this study the perfectly competitive solution, found 93 at one end of that spectrum, will be used as the basis for comparison. The actual market position would likely be more liberal as producer cooperatives exercise their market position and negotiation is utilized. Third, separating the area studied into only nine regions restricts analysis. One can not look at model generated regional results with making comparisons to a specific marketing order in mind. For example, not all shipments into or out of a given region must originate from or arrive at the region's market center. While in reality sales along order boundaries may represent a significant proportion of a region’s trade, when restricted to shipments between market centers, these sales may not continue. Given this understanding, the distributional jpmtterns generated by the model should be viewed as guides to changing flow patterns. Fourth, the vast shifts in Class I production levels «occurring under some scenarios do not reflect the local market phenomenon nor the ability for producer cooperatives 'ttb control markets and/or negotiate prices above the level gonerated by a theoretical model. Although this thesis docs not address these factors, they do exist and signifi- ccntly impact Class I marketings (USDA, January 1984). .Fifth, the greater the restrictions imposed, or the 191‘902- the alteration made to model parameters, the farther th. nodal is stretched and the less confidence one can have in “by given result. This is to say that given that supply 94 and demand schedules were generated from point elastic- ities, the further the solution is forced away from the equilibrium point at which the elasticities were applied, the less confidence one can have in the results. However, even though the model may not exactly parallel actual industry reaction to the conditions imposed, a sound theoretical indication can be garnered by comparing alternative scenario solutions to the BASE solution. Stating these obvious limitations is not meant to detract from the results generated nor the analysis submitted: rather, it is meant to serve as a reminder of the inherent limitations of such modeling. No perfect data sets exist nor is there a perfect theory through which to apply them. Given these limitations, the results and analysis from this study should be viewed as intended: as providing a useful indication of the possible impact of :certain technological and policy changes, given the .industry as modeled. 5-2 Base Run The model was run under 1985 data and market conditions I'lsth the results generated serving two key purposes. First, the results were directly compared to actual 1985 Iorket levels. This comparison served as a test of the 'Odtl's performance under normal conditions and it allowed Eh. lodel's specification to be recalibrated. This process 1.9 to the model as described in the previous chapter. 95 The model generated a base solution, BASE, with supply, demand and price levels very close to those calculated for 1985. Table 5.1 lists these levels and the percentage change between the actual and generated values. In general the model generated values remaining within II- 2 percent of actual levels; the exception being regions 4 and 8 for which the model generated prices 6.4 and 5.9 percent, respectively, less than the actual prices. While the model generated highly acceptable results statis- tically, one should not overlook the potential impact of even a minor change in variable levels. Table 5.1. Comparison Between Actual and Model Generated Supply Price and Ouantity Levels Price (Ilcwt) Ouantity (mil. cwt) Region Actual Model x Change Actual Model x Change 1 17.09 17.08 -0.06 20.21 20.20 -0.05 2 15.75 15.72 -0.19 27.85 27.81 -0.14 3 14.68 14.86 1.23 63.69 64.56 1.37 4 14.22 13.31 ~6.40 68.69 66.46 -3.25 5 14.68 14.83 1.02 52.59 53.03 0.84 6 14.34 14.06 -1.95 66.18 65.42 -1.15 7 14.79 14.57 -1.49 243.76 241.32 -1.00 8 13.80 12.98 -5.94 247.20 242.81 -1.78 9 13.88 13.67 -1.51 51.76 51.55 I0.41 The accuracy of distributional patterns for 1985 is not C- easily determined. A comparison between the source of OCCh region’s 1985 supply and the model's results match fairly well for ma3or shipment levels (over 10 million hufldrodweight). The exceptions being model shipments from z..Ql-Ol'es 6 and 9 to region 7, which did not actually occur 1" 1985, and the omission of shipments from region 4 to 96 region 8, which did occur. At lower minimum trade levels, the number of "wrong-way“ shipments (going south to north) increased. This is not surprising given the role of negotiation and the presence of overlap between states and marketing orders. The overlap between regions and marketing orders accounts for a large proportion of the difference between actual 1985 distribution patterns and those generated by the model. Region 4's shipments to region 8 are an example of this overlap effect. In 1985 Illinois shipped approx- imately 11 million hundredweight of milk to the Chicago Regional order. Presumedly, the maJority of this milk served the Chicago metropolitan area. Chicago is in Illinois which is part of region 4 but the Chicago Regional lorder is considered within region 8: hence, the large .shipments appearing to go north in 1985. When comparisons were made between actual interregional 'trade quantities and those generated by the model, concern cross over trade flow levels for region 7. During 1985, rcgion 7 imported approximately 9 million hundredweight, Yot, the model generated an import level of 35 million hundredweight. Either due to boundary overlap or not copturing some aspect of region 7's market environment, t‘910n 7 appears somewhat ”worse-off" in the model. This 1. likely carried through all runs. It was not apparent ”’3‘":- form of ad3ustment should be made based on theory and “PSQrstanding that market: hence, rather than guessing at 97 how the specification should be altered, if at all, region 7 was left as is. Interpretation of results tied to that region should be made with greater care. In all, however, the generated distributional patterns do capture the maJor flows and will provide useful insights as to what the true distributional patterns would be. Figure 5.3 provides an illustration of distributional patterns under the BASE run. .0‘ Figure 5.3. Distributional pattern under 1985 market conditions, (BASE) The second purpose for the 1985 run was to provide a bonch mark upon which solutions from alternative runs could 5‘ compared. Given the models good performance relative to actual 1985 levels, one can feel confident that the 1985 bg‘. Qolution adequately serves as a basis for comparison. 98 5.3 Impact of R0 Filtration The regional impact of RO filtration on the fluid milk industry is determined by comparing the model generated RO solutions to the BASE solution. As discussed in the previous chapter, the exact cost of full scale operation of an R0 filtration unit for whole milk is unknown. To determine the range of possible impacts of R0 filtration on the industry, three widely varying cost estimates were employed, 8.30, 0.90, and 81.75 per cwt. In addition to creating a range of possible impacts, such varied cost levels provide an indication of how sensitive the model is to increasing the fixed cost component of the transpor- tation cost function. 5.3.1 Applying RO Filtration at 0.30/cwt When use of R0 filtration is priced at 0.30/cwt, distribution patterns, production levels and Class I allocations change significantly. Table 5.2 provides a reference to the percentage changes occurring under this scenario, BR03. It should be noted that in regard to changes in exporter revenues, the focus should remain on total rather than Class I revenues. The reason for this being that producers receive a blend price, not Just the Class I price. Total revenues were calculated as a weighted average of Class I and Class III revenues. Across all regions Class I sales increase negligibly, .56 percent, but regionally, the impact is highly skewed. 99 The availability of essentially half-priced transportation allows region 8 to capture significant gains due to its relatively lower costs of production and Class I differ- ential. While total exports remain nearly the same, Class I exports in region 8 increase 824 percent as it becomes the sole supplier to regions 1, 2, and 4 and the primary supplier to region 7. Table 5.2. Market Impact Under 1985 Conditions with R0 Applied at 0.30/cwt, (BRO3) Regions Variable All 1 2 7 8 A 9 (Numbers are as a percentage change of BASE values) Exports Total -10.74 -7.44 0.02 0.01 -0.14 Class I 0.56 -100.00 -100.00 -100.00 824.12 -0.15 Class III 792.13 824.17 52.51 -87.50 0.00 Price -6.95 -13.24 -9.25 0.03 0.03 -0.51 Revenues Total -6.44 -28.84 -7.52 8.02 0.60 Class I -6.42 -100.00 -100.00 -100.00 824.42 -0.66 Imports Total 1.78 1.29 0.62 .00 0.04 Costs -6.26 -15.44 -12.25 -7.69 0.03 -0.49 No other region experienced a gain in sales. In fact, regions 1, 2, 3 and 7 no longer compete in the Class I market, to an accountable level, and region 4 experiences a 93 percent drop in Class I sales. The BROS distri- butional pattern is illustrated in Figure 5.4. A visual comparison of the BASE and BRO3 distribution pattern demonstrates how the shift in distribution favors region 8. 100 Figure 5.4. Distributional pattern under $985 market conditions with R0 applied at 8.30/cwt, (BROS) In terms of new Grade A production levels, it is unclear exactly how much milk would continue to be produced solely for Class III use within regions no longer competitive in the Class I market. It is fairly safe to say that some localized production for Class I use would continue. Production for Class III would depend upon local demand and the producer's ability to remain in operation at the government supported Class III price. As specified, the modal dumps excess Class I supply into the Class III sink. With a government set minimum Class III price of 11.78, and aggregate regional coats of production in regions 1, 2, 3, 4, and 7 above that level, the livelihood in those regions is uncertain. However, 101 with the model solutions being generated from annual data and supply and demand quantities listed in millions of hundredweight, no steadfast assessment should be made. This dramatic shift in distribution is tied to the alteration in regional export pricea. Across all regions a 6.9 percent decrease in export price occurs. Region 1 experiences the largest drop, over 13 percent. The resulting withdrawal of region 1 from the Class I market suggests that a 13 percent fall in Class I price would not only put producers in that region at a significant market disadvantage but also it may force many out of production. The possibility of producers dropping out of the market is further strengthened by the impact on regional Class I revenues. Producer Class I revenues are dependent upon total sales and market prices. Regions which fall out of the Class I market will see a 100 percent decline in Class I revenues: likewise, region 8 enJoys a revenue increase on the same proportion as sales, 824 percent. The more representative total revenues do not shift as dramatically. Region 8, gains 8 percent, while all other regions lose in total revenues. The actual decrease for regions 1, 2, 4 and 7 remains dependent on how many producers can remain in operation at the Class III price. The large regional boundaries, quantity units and the short-run characteristic of the model make determination impossible. 102 On the consumer side, the import price falls 6.23 percent. Even with this substantial fall in price, the inelastic nature of milk demand leads to a mere .56 percent increase in consumption. Regionally, consumers in the south en3oy the greatest savings as import prices fall up to 15 percent. 5.3.2 Applying RO Filtration at 0.90/cwt When the cost of applying RO filtration is tripled to 8.90/cwt, model results are again significantly altered, albeit not as dramatically. Figure 5.5 illustrates the new interregional flow pattern and Table 5.3 presents the percentage changes associated with this run, BRO9. 1 a" Figure 5.5. Distributional pattern under 1985 market conditions with R0 applied at 8.90/cwt, (BRO9) 103 Table 5.3. Market Impact Under 1985 Conditions with R0 Applied at 0.90/cwt, (BRO9) Regions Variable All 1 2 7 8 9 (Numbers are as a percentage change of BASE values) Exports Total -9.96 -7.25 0.01 0.01 -0.14 Class I 0.35 -11.06 -8.06 -100.00 436.59 -0.15 Class III 0.00 0.00 52.20 -46.35 0.00 Price -5.11 -12.27 -9.01 0.02 0.02 -0.52 Revenues Total -20.41 -15.09 -7.52 4.25 -.061 Class I -4.78 -21.98 -16.35 -100.00 436.72 -.067 Imports Total 1.39 0.90 0.30 0.00 0.50 Cost -3.86 -12.02 -8.61 -3.77 0.02 -0.50 The stair-step effect found under the base solution disappears. Region 8 again captures markets in the two southernmost regions shipping 15 percent of its Class I sales, which corresponds to 51 percent and 45 percent of region 1 and 2’s demand, respectively. Additionally, region 8 supplies 70 percent of region 7’s Class I require- ment. Region 7 is the only region under this scenario which ceases production at the new equilibrium price level. Region 4 experiences a shift of up to 13 percent of Class I production either out of production or into the Class III market. Across all regions, the Class I price falls 5.11 percent. As expected, the greatest impact occurs in regions 1 and 2 where prices decline 12.27 and 9.01 percent. In contrast, regions 7 and 9 see no significant change in their Class I export price. 104 The inevitable impact of declining producer prices in the absence of an equivalent rise in sales leads to a drop in revenues. Revenues from Class I sales fall 4.78 percent across all regions. Excluding region 7, regions 1 and 2 record the largest falls, 21.98 percent and 16.35 percent respectively. As in the previous case, only producers in region 8 appear to capture gains in Class I sales and Class I revenues (436 percent each). Consumers remain significant gainers under this scan- ario. Across all regions the import price falls 3.85 percent, ranging from a high of 12 percent in region 1 to virtually no change in regions 8 and 9. The overall impact on demand of a lower Class I price remains negligible. 5.3.3 Applying RO Filtration at 81.75/cwt When R0 filtration is priced at 01.75/cwt the market impact is relatively minimal on the whole, as indicated by Table 5.4 and Figure 5.6. Intuitively this makes sense given a one-way break even mileage of 772 miles under this scenario. With export pricea falling 1.3 percent, total Class I exports remain unchanged and the regional dist- ribution pattern begins to resemble that of the BASE solution. 105 Table 5.4. Market Impact Under 1985 Market Conditions with R0 applied at 81.75/cwt, (BR0175) Regions Variable All 1 2 7 8 9 (Numbers are as a percentage change of BASE values) Exports Total -5.93 -2.91 0.01 0.00 0.00 Class I 0.12 -6.58 -3.24 -1.84 62.64 0.00 Class III 0.00 0.00 0.01 0.01 0.01 Price -1.30 -7.31 -3.62 0.01 0.01 0.01 Revenues Total -12.45 -6.22 -0.13 0.61 0.01 Class I -1.19 -13.41 -6.74 -1.83 62.62 0.01 Imports Total 0.82 0.36 0.00 0.00 0.00 Cost -1.21 -7.15 -3.46 0.01 0.01 0.01 4'? 4 'N‘ \ I. Figure 5.6. Diatributional pattern under 1985 market conditions with R0 applied at 01.75/cwt, (BR0175) 106 Of significant change is the opening of new markets for region 8. As in the previous two cases region 8 exports to regions 1 and 2 when RO filtration is adopted. These shipments replace those made by regions 2, 4, and 6 under the BASE run. Also of importance is the re-entry of region 7 into the Class I market. In fact, region 7 experiences only a 1.84 percent decline in Class I sales. This suggests that region 7's sensitivity to the change in market variables, arising when RO is applied, is more likely a result of its relatively high Class I differential than due to its relative cost of production. Reductions in Class I sales occur as in the previous cases. Regions 1 (6.57 percent) and 2 (3.24 percent) experience significant declines while once again region 8 gains substantially (62 percent). In addition, export prices fall by 7.3 percent in region 1 and 3.6 percent in region 2 with all other regions experiencing insignificant changes. Total producer revenues for Class I sales fall 1.19 percent with the brunt of this decline again being borne by producers in regions 1 and 2 (13.4 and 6.7 percent). In contrast, region 8 en3oys increased Class I revenues of more than 62 percent. It is again, however, difficult to determine the impact on total revenues. 107 5.3.4 Interpretation of Annual R0 Results Within the perfectly competitive, short-run, spatial equilibrium framework applied through this model, RO filtration would provide no real benefit to producers as a whole, even when priced at its lowest level. The net effect of full scale adoption of R0 filtration is a significant loss in total revenues to the industry, under all RO cost scenarios. In fact, only region 8 stands to gain and only under the R0 equals 8.30/cwt and 8.90/cwt scenarios. Additionally, while the possibility exists for some regions to capture new markets, without significantly increased levels of demand, any market gained by one region will represent a loss to another. Such is the case within the three scenarios described in this section. It should be noted that the results are unclear as to exactly how much milk would continue to be produced in regions no longer competitive in the Class I market. It is fairly safe to say that some very localized production of milk for Class I and Class III use would continue to the extent which local demand warrants and producers can remain viable at the government supported Class III price. Hence, these vast shifts in Class I production levels do not reflect local market phenomenon nor the ability for producer cooperatives to control markets and/or negotiate prices above the level generated by a theoretical equilib- rium model. This caveat holds true for all the model generated solutions in this study. 108 Considering the large number of RO facilities which would likely be operated by producer cooperatives, it is unforeseeable that those cooperatives would voluntarily adopt any technology which would result in lost revenues for their producers. However, recognizing the important role of negotiation and the acquired market power which exists within certain markets, it may be possible for some of the benefit gained by consumers in the theoretical case to be usurped by producers. Under BASE scenario conditions with R0 filtration priced at 8.90/cwt, this represents a possible average gain of around 4 percent. In an industry where profit margins are slim and producers are struggling to remain viable, this represents a significant increase. How much of that gain could be negotiated away from consumers and how it would be distributed regionally among producers is not clear. In sum, the future for R0 filtration under BASE con- ditions does not look politically promising. In the theoretical case, sufficient benefits must be created to compensate the losers. Given the lopsided nature of the costs and benefits generated by the model and, more importantly, considering the underlying costs associated with the necessary policy changes and transfers of bene- fits, it is doubtful that RO filtration would find use under these conditions. 109 5.4 September Conditiona When modeling the Class I milk market using annual data, the inherently strong seasonality of supply and demand is lost through averaging. In attempting to determine the market feasibility of a technology such as RO filtration, which directly impacts the transportation component of fluid milk shipments, it would then seem of obvious interest to analyze the technology’s impact during the time when shipments are naturally highest. Running the model for the month of September serves this purpose; demand is near its peak, supply is approaching its trough, and inter- regional shipments are at their annual high. If this shipment hypothesis is correct, one would expect to see the benefits of R0 filtration to be greatest during September. Once again, however, this modeling exercise should be viewed given the limitations of the model. Specifically, the following points should be considered. First, the model was designed for, built upon and calibrated for annual data. To define short-run as one month stretches the accuracy of the underlying short-run supply and demand elasticities. However, given the model’s relative insensitivity to elasticity changes, as discussed in a later section, this should not present a significant problem and no attempt has been made to re-specify the existing annual model for monthly data. Secondly, what can be easily averaged out on an annual basis can not necessarily be disregarded on a monthly 110 basis. For example, the existence of sustained call provisions in New York made it necessary to force a set quantity of region 7’s September supply to the Class I market. This should not create any problems. Third, no ad3ustments are made to the estimated costs of RO filtration. It is assumed, therefore, that RO filtra- tion is employed during September at a cost analogous to that used under annual conditiona. Given that September test runs were made for the 8.30 to 81.75/cwt range of RO costs, the true average cost for any number of use conditions is presumably covered. While these points of caution exist, none is over- whelming. It is believed that the results presented in this section do serve the purpose for which they were intended: to gain insight into the possible impact of R0 filtration on distribution flows, prices and revenues during September market conditions. 5.4.1 September Base Run Under the base September run (SEPT), the model gener- ated regional supply and demand levels averaged within Il-1.4 percent of actual levela. While prices generated are higher than actual prices, the more important relative alignment of prices regionally remains intact. Distributional flow patterns for SEPT do not reveal such in and of themselves (Figure 5.7). Comparison to the actual pattern is not possible given the lack of 111 \. 1 a, . Figure 5.7. Distributional pattern under September market conditions, no RO applied, (SEPT) appropriate data. It is worth noting that while regions 4 and 8 do not show shipments for September in addition to their annual average, regions 6 and 9 have each increased their export markets. This satisfies the belief that regions with large excess annual Class I supplies are more likely to make September shipments to regions whose annual supply and demand quantities are more in line -- suggesting September supply shortages. All in all, the base September run indicates that substituting September data into the fully specified annual model does provide a reasonably accurate solution upon which comparisons can be made. Runs were generated for the 112 three alternative RO scenarios (SR03, SR09, and SR0175) with the percentage changes listed in Tables 5.5, 5.6 and 5.7. The following discussion of these results focuses on the RO filtration scenario in which cost is set at 0.90/cwt, SR09. Table 5.5. Market Impact Under September Conditions with R0 Applied at 0.30/cwt, (SRO3) Regions Variable All 1 2 7 8 9 (Numbers are as a percentage change of SEPT values) Exports Total -10.14 -7.23 0.03 0.01 0.01 Class I 0.64 -100.00 -100.00 0.64 482.91 -31.44 Class III 782.05 817.75 -0.58 -52.37 58.42 Price -5.22 -12.54 -9.00 0.04 0.04 0.04 Revenues Total -32.50 -25.04 0.10 2.11 -1.49 Class I -4.62 -100.00 -100.00 0.68 483.16 -31.41 Imports Total 1.65 1.23 0.66 0.00 0.00 Costs -4.85 -14.53 -11.79 -0.38 0.04 0.04 Table 5.6. Market Impact Under September Conditions with R0 Applied at 8.90/cwt, (SR09) Regions Variable All 1 2 7 8 9 (Numbers are as a percentage change of SEPT values) Exports . Total -9.23 -6.71 0.02 0.01 0.01 Class I 0.42 -10.28 -7.46 0.64 76.62 -31.44 Class III 0.00 0.00 -0.60 -8.30 58.41 Price -3.07 -11.42 -8.36 0.02 0.02 0.02 Revenues Total -18.95 -13.95 0.08 0.34 -1.50 Class I -2.66 -20.52 -15.19 0.67 76.66 -31.42 Imports Total 1.27 0.83 0.47 0.00 0.00 Costs -3.01 -11.17 -7.96 0.03 0.02 0.02 113 Table 5.7. Market Impact Under September Conditions with R0 Applied at 81.75/cwt, (SR0175) Regions Variable All 1 2 7 8 9 (Numbers are as a percentage change of SEPT values) Exports Total -4.98 -2.14 0.01 0.00 0.00 Class I 0.23 -5.55 -2.38 0.64 60.96 -30.89 Class III 0.00 0.00 -0.62 -6.61 57.38 Price -1.17 -6.17 -2.66 0.01 0.01 0.01 Revenues Total -10.50 -4.57 0.06 0.27 -1.49 Class I -0.94 -11.37 -4.98 0.65 60.98 -30.89 Imports Total 0.69 0.26 0.47 0.00 0.00 Cost -1.07 -6.03 -2.54 0.01 0.01 0.01 5.4.2 Application of RO Filtration The impact of R0 filtration on fluid milk marketing under September market conditions is substantial. All scenarios follow the percentage change pattern outlined under the annual scenarios. This is to say that when RO filtration is priced at 8.30/cwt (SRO3) the most signif- icant change occurred: when priced at 81.75/cwt (SR0175), the impact was substantially less. For example, under SRO3 export prices fell by 5.2 percent across all regions with the largest decrease hitting southern regions and with greatest gain in Class I sales captured by region 8 (483 percent). The significance of these shifts diminishes under SR0175: the market wide export price increases 1.2 percent and region 8’s Class I exports increase 61 percent. When R0 is introduced at a cost of 0.90/cwt, the impact on the industry is significant. The application of I 114 R0 filtration leads to a 3 percent overall decrease in the export price. As expected, the burden of this decrease to producers and the benefit to consumers falls heaviest in the two southernmost regions. Total revenue loss for regions 1 and 2 is 19 and 14 percent respectively. Comparison of Figures 5.7 (SEPT) and 5.8 (SR09) illus- trates how the distribution pattern of Class I milk shifts. Specifically, region 8 becomes a competitive supplier of Class I milk to regions 2 and 5, and region 4 ships south to region 3. In contrast to these market gains, however, several regions find markets once open to them during September, now served by alternate sources. For example, both region 6 and 9 lose two export markets. \I* E\m - l ‘_ “ \ 2 I \ ”‘h a“ Figure 5.8. Distributional pattern under September market conditions with R0 Appliad at 8.90/cwt, (SR09) \. 115 The magnitude of this shift presents itself in Table 5.6. It is interesting to see how application of R0 filtration reduces the price gap once existing between regions 8 and 9 due to distance. Figure 5.8 suggests that region 9 may loose its market directly to region 8. This may suggest that under this scenario, the application of RO filtration eliminates enough of the mileage disadvantage faced by region 8 to allow it to capitalize on its relative competitive advantage in milk production. Furthermore, while regions 7, 8, and 9 do not see a significant change in their export price, only region 9 suffers a fall in Class I exports, 31 percent, as its exports shift from Class I markets to Class III. If region 9’s total produc- tion does remain unchanged, the total revenue loss is limited to 1.5 percent. In sum, given that the net impact to all producers remains negative, even under September market conditions, R0 technology would not likely be adopted under a market environment described by the model. It is worth noting however, that the ”negative" impact to producers under September conditions is less than that found under the annual case. From the point of view of an "imperfect" market, this would suggest that for any constant percentage of benefits negotiated at an annual level, a higher percent of the consumer's gains could possibly be transferred to producers during September. 116 5.5 Cut in 000 Purchases At present the dairy industry is operating during an era in which government is looking for areas to trim its budget and agricultural enterprises are buckling under both heavy debt and a competitively fueled push for increased effic- iency. Under this environment, government expenditures on manufactured dairy products have become highly visible and open to increased public criticism. Without 3udging the merits of either U.S. dairy policy or criticisms of CCC operations, the model was run under a scenario of reduced CCC purchases. The volume of 1985 CCC purchases was roughly equivalent to 13.2 billion pounds. An estimated national decrease of 8.2 billion pounds would be necessary to achieve a purchase level approximately at equilibrium with government demand. It was determined that approximately 63 percent of 1985 CCC purchases came from the area covered in this study (DMS, 1985, Table 7). Hence, 4.44 billion pounds of the reduction must be met within the model. The cut in CCC purchases was incorporated into the model by setting an upper limit on Class III imports. Subtrac- ting 44.4 million hundredweight from the total quantity of Class III shipments in the BASE equilibrium aolution produced the proper Class III import quota level. Such a quota should, and in fact does, force more milk onto the Class I market. This, in turn, precipitates a price decrease across all regions. Results from the reduced CCC 117 purchases run (CCC), presented in Table 5.8 support this chain of events. Table 5.8. Market Impact Under 1985 Conditions with CCC Purchases Reduced, (CCC) Regions Variable All 1 2 7 8 9 (Numbers are as a percentage change of BASE values) Exports Total -6.92 -6.88 -6.77 ~3.23 -2.83 C1... I 0.70 -7a69 -7e65 24e20 Oe6° -3a14 Price -8.67 -8.53 -8.55 -10.00 -11.22 -10.66 Revenues TOt.l -15e34 -15a30 -15e09 -15e04 -13a37 Class I -8.04 -15.57 -15.54 11.78 -10.69 -13.47 Imports Total 0.96 0.86 0.77 0.60 0.82 COIL -8e23 -8e35 -8s17 -9a53 -11e22 -10s26 Under limited CCC purchases, the weighted average price of exports falls 8.7 percent and the weighted average import price falls 8.2 percent. The new Class III price drops approximately 12.4 percent to a level of 010.32. Of the 4.44 million hundredweight removed from Class III use within the model, additional Class I sales. represents a decrease in total supply. The remaining 93.3 percent only 6.7 percent is absorbed as This would follow the hypothesis that under a significant price decreaae, producers operating near the margin will be forced to discontinue production. relatively high cost of production regions would be Regionally, those producers in the expected to absorb the greatest impact of the reduction, which they appear to do. Regions 1 and 2 decrease total 118 supply by about 6.9 percent each while regions 8 and 9 see total supply fall by 3.2 and 2.8 percent respectively. In regard to lost sales, all of the southern regions’ loss comes from Class I sales. This is also the case with region 9. Only region 7 sees significant increases in Class I sales, 24 percent: but, no region experiences an increase in total revenues. Across all regions, producer revenues fall an estimated 14 percent. The distribution of this lost revenue, however, is not as clear cut north to south as under previous scenarios. Specifically, regions 7 and 8 Join regions 1 and 2 in suffering above average losses, over 15 percent, while region 9 suffers less of a loss, 13 percent. The new distribution pattern is presented in Figure 5.9. Region 4 begins shipping a small amount of its supply to region 3, region 2 looses its region 1 Class I market and both region 6 and 9 gain markets. It is the addition of new markets which helps to limit revenue losses in these regions to below the industry average. The impact to the consumer of lower CCC purchases and, hence, a lower Class I support price, is examined by looking at the new import price. Across all regions the price of imports falls 8.2 percent: however, in contrast to previous scenarios, no longer is the greatest drop in prices, either import or export, found in the south. Rather, the greatest percentage change is found in region 8, 11.2 percent. 119 ~ 4\ _ aw - ...‘ Figure 5.9. Distribution pattern under 1985 market conditions with CCC purchases reduced, (CCC) 5.5.1 Applying R0 Filtration at 8.90/cwt The industry impact of both R0 filtration and a signi- ficant drop in the M-W price is not uniform. Results presented in Table 5.9 and Figure 5.10 (from run CCCR09) indicate that application of 8.90/cwt RO filtration benefits regions with relatively low Class I differen- tials. These are some of the same regions which suffered . the greatest relative burden from the initial decrease in M-W price. For example, application of RO filtration cuts the loss in total revenues experienced by producers in regions 8 and 9 from approximately 15 and 13.4 percent to Tot Clo Clo Pri Rev Tot Cle II] Toi 120 Table 5.9. Market Impact Under 1985, Reduced CCC Con- ditions with R0 Applied at 8.90/cwt, (CCCRO9) Regions Variable All 1 2 7 8 9 (Numbers are as a percentage change of BASE values) Exports T°t.1 -14e75 -12e42 -4a67 -2e23 -2e10 01". I 0a88 -16a39 -13a80 -100eoo 475s21 -2e46 Class III 0.00 0.00 45.35 -52.93 0.00 Price -11.90 -18.19 -15.44 -6.90 -7.75 -7.90 Revenues TOtal -30a22 -25a97 -22a76 -7e49 -10a25 Class I -11.12 -31.60 -27.11 -100.00 430.64 -10.17 Imports . Total 2.06 1.55 0.84 0.41 0.61 COIL -10a02 -17a80 -14a74 -10s38 -7e75 -7e61 4 \‘W N ‘ I ‘ 3 \ t Figure 5.10. (CCCRO9) a" Distribution pattern under 1985, reduced CCC purchases market conditions with R0 applied at 8.90/cwt, 7.5 o: lenti; corre perce 8 on: pure] burd. and R0 1 difi con] ther Pre log fur CC: ou. oe; Dr Uh. Ql l"l 121 7.5 and 10.3 percent respectively. (When viewed as imple- menting R0 filtration after the cut in M-W price this corresponds to an actual increase in revenue of 8.9 and 3.6 percent for a price increase of 4 and 3 percent in regions 8 and 9 respectively). The net effect of a cut in CCC purchases and application of R0 filtration is for the burden of a lower M-W price to shift away from regions 8 and 9 towards regions 1, 2, and 7. In contrast to the benefits enJoyed by regions 8 and 9, RO filtration causes producers in the relatively high differential regions, such as 1, 2 and 7, to sustain compound negative impacts. First, the cut in CCC causes them to suffer substantial revenue loses, as discussed previously. Second, R0 filtration forces an additional lowering of prices and Class I sales, forcing revenuea even further down in these relatively high cost of production regions. 5.5.2 Interpretations of Results The market impact of a 8.2 billion pound reduction in CCC purchases is significant. As CCC purchases decrease supplies previously allocated to government purchases now serve to flood the Grade A market. This, in turn, forces prices down. With the M-W price set at 010.32, it is unclear how many producers can remain competitive. Clearly all dairy operations producing at costs above that level will be forced out of production over time. This is true 122 both within and across regions. The exact number of producers withdrawing from the market is unclear. The model generates a short-run solution while reduced CCC purchases would have a long-run effect. The results suggest that in the face of a significant reduction of the M-W price, to maintain regional production at levels analogous to current levels, the Class I differ- entials would have to be increased. Interestingly, this appears to have happened with the 1985 Food Security Act. The Class I differentials was increased while the M-W price fall nearly two dollars over the previous five years. The net effect of a decrease in CCC purchases and the adoption of R0 filtration is to leave relatively low cost of production regions relatively better off than other regions. Specifically, regions 8 and 9 may capitalize on both their comparative advantage in production and their relatively low Class I differentials. Once again, policy changes, i.e. ad3usting the Class I differentials, would be necessary in order to preserve the status quo. 5.6 Altering the Class I Differentials One of the barriers commonly cited to full scale adoption of bulk reducing technologies is the Class I differential system. Given this, it should prove interes- ting to explore the impact which both removing and realig- ning these differentials would have on model solutions. This section discusses each of these policy scenarios. 123 5.6.1 Removal of Class I Differentials With Class I differentials removed from the price linkage function, the base model was rerun. The new solution, ND, was than compared to the original BASE solution. Figure 5.11 illustrates the shifts in shipment patterns which occur while Table 5.10 highlights the market changes which lead to these shifts. Figure 5.11. Distribution patterns under 1985 market conditions with Class I differentials removed, (ND) 124 Table 5.10. Market Impact Under 1985 Market Conditions with Class I Differentials Removed, (ND) Regions Variable All 1 2 7 8 9 (Numbers are as a percentage change of BASE values) Exports Total 11.91 6.66 -0.24 -0.17 -0.67 Class I 0.15 13.23 7.40 42.13 0.03 -50.01 Class III 0.00 0.00 -22.47 -0.19 441.42 Price -13.65 -5.42 -8.59 -19.15 -9.24 -13.79 Revenues Total 6.59 -1.68 -7.75 -1.14 -13.17 Class I -13.78 7.09 -1.82 14.92 -9.21 -56.91 Imports Total -1.70 -0.92 -0.06 0.03 0.17 Cost -13.55 -5.31 -8.20 -18.25 -9.24 -13.27 An interesting impact under this scenario is on distri- bution patterns. One purpose of setting differentiala at levels increasing with distance from the base pricing point was to ensure a steady supply of milk at the local level. This was accomplished by the differentials ability to support production in high cost of production regions via raising the minimum price. If the differentials are properly set, their relative level should off-set transpor- tation costs. Given this, removal of differentials should have little effect. If they are improperly aligned, one would expect that their removal would alter prices and increase interregional shipments as the market realigns itself to costs of production. When the differentials were removed from the model, interregional distribution of Class I milk actually decreases. The Class I price across all regions falls 125 substantially (over 13 percent) with the relative regional export price roughly increasing the larger the region's cost of production is to its Class I differential. Corresponding to the price fall is a 13 percent drop in Class I revenues. Perhaps surprisingly, the greatest weight of this burden appears to be borne by producers in region 9. In contrast, region 1 sees the lowest fall in export price. Furthermore, region 1 actually increases its total value of exports by increasing production by a greater percentage than the fall in price. The only sure gainers are the importers of Class I milk, the consumers. An additional run was made under which industry wide application of 0.90/cwt R0 filtration was instituted after the removal of Class I differentials. While prices did change, no shift in the distribution pattern occurred. This is as would be expected when an unconstrained market in equilibrium receives an equal decrease in transportation costs across all regions. If these results are at all indicative of how the industry would actually react under elimination of Class I differentiala, it appears that the 1985 differentials not only serve to maintain production at the local level but they also provide the incentive behind a large proportion of interregional shipments. As long as these shipments are made to help balance local Class I supply with Class I demand, they serve a beneficial role. However, it appears 126 that in some regions they may actually decrease the proportion of Grade A milk going to Class I use due to the effective subsidization of exports from other regions. 5.6.2 Realigning Class I Differentials Testing the impact of alternative Class I differentials became necessary with passage of the 1985 Food Security Act. This legislation set new Class I differentials for all regions. On average these differentials, increase at an increasing rate from the base pricing point. To determine how model results would be altered under this legislation, new regional differentials were calculated based on the 1986 levels and incorporated into the BASE model. It quickly became evident that relative changes in regional differentials have a significant impact on model results. 5.6.2a Impact on Base Run Initial distribution patterns, as illustrated in Figure 5.12, show a rather mild impact of the new differentiala (run D86). Region 6 looses one market and region 8 replaces region 4 as an exporter to region 2. In terms of regional Class I prices and revenues, however, the impact is significant. Table 5.11 provides a reference to the degree of change resulting from the new differentials. 127 he. 1 ..0' Figure 5.12. Distribution patterns under 1985 market conditions and 1986 Class I differentials, (D86) Table 5.11. Market Impact Under 1985 Market Conditions and 1986 Class I Differentials, (D86) Regions Variable All 1 2 7 8 9 (Numbers are as a percentage change of BASE values) Exports Total -4.30 -2.48 1.20 0.24 0.72 Class I 0.00 -4.78 -2.75 -3.13 56.02 0.80 Class III 0.00 0.00 3.48 -5.68 0.00 Price 0.18 -5.29 -3.08 1.78 0.85 2.70 Revenues Total -9.11 -5.30 1.56 0.92 3.21 Class I 0.18 -9.82 -5.75 -l.41 57.35 3.52 Imports Total 0.60 0.31 -0.14 -0.05 -0.21 Cost 0.15 -5.18 -2.94 1.69 0.85 2.60 128 Across all regions the export price and revenue from Class I sales remain unchanged. Regionally, variation exists. The Class I export price in region 9 shows the greatest gain, 2.7 percent, with region 1 receiving the greatest loss, 5.3 percent. In terms of revenues, region 8 stands to gain substantially with an increase of 57.3 percent: region 9 follows with a 3.5 percent gain. On the consumer side, regions 7 and 9 appear to feel the greatest impact as their import prices increase 1.7 and 2.6 percent respectively. The changes described above all indicate that the disproportional increase in Class I differentials tends to favor producers in regions 8 and 9. This unbalanced effect is further illustrated by model results when R0 filtration is introduced. Highlights of these results are discussed below. 5.6.2b Applying RO Filtration at 0.90/cwt The industry wide application of RO filtration, given 1986 differential levels, appears to emphasize the apparent imbalance caused by the new Class I differentials. Percentage changes between this scenario, D86RO9, and 086 are listed in Table 5.12. Figure 5.13 portrays a radical change in distribution patterns both relative to the BASE (Figure 5.3) and BRO9 (Figure 5.5) patterns. The most obvious change is the new role of region 8. It becomes a key supplier of Class I milk to three regions and the sole Close 1 Close 1 Price ROVOnue Totol Close 1 I‘Porti Totol Coot \ Table 5.12. 129 Market Impact Under 1985 Conditions and 1986 Class I Differentials with R0 Applied at 0.90/cwt, (D86R09) Regions Variable All 1 2 7 8 9 (Numbers are as a percentage change of D86 values) Exports 101.181 -4a87 ‘3e27 1 e20 1 e26 0a87 Class I 0.19 -100.00 -100.00 -100.00 848.84 0.96 Class III 850.77 865.82 54.31 -88.74 0.00 Price -4.01 -6.01 -4.06 1.78 4.40 3.27 Revenues Total -32.27 -25.64 -6.42 13.84 3.89 Class I -3.84 -100.00 -100.00 -100.00 890.59 4.26 Imports Total 1.76 0.99 0.09 -0.24 -o.25 Cost -1.80 -15.24 -9.47 -1.11 4.40 3.15 \\ .‘---~ 4 '\ ‘ \ A ___ 5 ' In 3 / ‘\\ 2 .0“ Figure 5.13. Distribution pattern under 1985 market conditions and 1986 Class I differentials with R0 applied at 0.90/cwt, (D86R09) 130 supplier to two. Region 8’s Class I sales increase dramatically, over 500 percent, as regions 1, 2, 3 and 7 fall out of the Class I market. Region 9 also benefits significantly form the adoption of RO filtration under these market conditions. It’s level of total revenue increases nearly 4 percent as market price raises. 5.6.2c Interpretation of Results The results from these runs tend to demonstrate two important points. First is the ability of a relative change in the differential level to alter the existing balance within the industry. Any marketing activity operating near the margin is easily influenced by such a change. Within a spatial equilibrium context, any sig- nificant alteration to one region's market will be felt across all regions. Such is the case with implementation of new Class I differentials. As relative prices change, marketings and revenues change. Second is the choice presented by the availability of R0 filtration: to allow increases'in some forms of market efficiency versus maintaining the status quo. The intro- duction of RO filtration does allow for increased effic- iency in the sense of comparative advantage and trade theory: however, the burdens and benefits of transition to such a market do not fall evenly. There is no Pareto optimal solution. 131 5.7 Additional Runs After running the model under a range of alternative scenarios, two main areas of question remain to be discus- sed. Specifically, how does the model react to a sign- ificant increase in the cost of transportation and, results change when the underlying supply and demand elasticities are altered. Each of these areas will how do be discussed below with comparisons made where feasible. 5.7.1 Increasing Transportation Costs In recent years the transportation industry has seen significant increases in operation costs. In an effort to yield insight into the variability of model results under conditions of significantly increased shipment coats, each component within the transportation cost function was increased by 50 percent. The following set of transpor- tation cost (TC) functions result for one-way mileage: Short Haul: TC I .25008 I No R0 applied: TC I .24737 0 Cost of RO I 0.30/cwt: TC I .42368 9 Cost of R0 I 0.90/cwt: TC I1.02368 0 Cost of R0 I01.75/cwt: TC I1.87368 9 Table 5.13 presents the revised long haul distances and Figure 5.14 illustrates the .00998M .00648H .00324H .00324H .00324H break even cost versus (5.7a) (5.7b) (5.7c) (4.3d) (4.3a) mileage relationship under these increased costs and the original transportation costs. Using these new functions, 132 the BASE and BRO9 scenarios were regenerated as T02 and TC2RO9. Results from these runs are presented in the following section. Table 5.13. Long Haul Break Even Cost Distances Under 50 Percent Increased Transportation Costs Cost of Break Even Cost Function RO Distance Sequence M < 100 (5.7a) None ---- M > 100 (5.7b) M < 100 (5.7a) 0.30/cwt 54 miles M 100 (5.7c) M < 100 (5.7a) 100 > M > 240 (5.7b) 0.90/cwt 240 miles M > 240 (5.7d) M < 100 (5.7a) 100 > M > 502 (5.7b) 01.75/cwt 502 miles M > 502 (5.7a) OM) mi 8001 g 800-4 4 g 400 2 300-1 200-1 100 0| I I f I T 1 T I r l 1 I T I 0.3 0.6 0.7 0.0 1 .1 1.3 1 .6 1 .7 mnueflROmeMme u Numu I manque Figure 5.14. Break even mileage of RO filtration under unad3usted and fifty percent increased transportation costs 133 5.7.1a Impact on Base Run One would expect a 50 percent increase in transportation costs to result in both a reduction in interregional Class I milk shipments and an increase in intraregional Class I sales, especially within newly ”isolated" markets. Indeed, this is the case. Figure 5.15 illustrates an obvious reduction in shipments when compared to Figure 5.3. For example, region 1 becomes self sufficient while region 2 draws its additional supply from a closer source (region 5 versus region 4). .0 J Figure 5.15. Distribution pattern under 1985 market conditions and fifty percent increased transportation costs, (TC2) 134 The percentage change comparisons shown in Table 5.14 also uphold the solutions compatibility with theory. The increased per mile transportation cost restricts the ability of lower relative cost of production regions from capitalizing on their combined comparative advantage and lower relative Class I differentials. Region 8 appears to be hardest hit suffering a 50 percent decline in Class I sales. This corresponds to a loss in total revenues of over 7.7 percent. The biggest gainer is region 7 (total revenues up 3.9 percent) which no longer finds itself losing markets to more distant regions because of its relatively high Class I differential. Table 5.14. Market Impact Under 1985 Conditions and Fifty Percent Increased Transportation Costs, (TC2) Regions Variable All 1 2 7 8 9 (Numbers are as a percentage change of BASE values) Exports Total 7.35 3.05 0.00 0.00 -0.55 Class I -0.23 8.16 3.39 51.49 -0.11 -50.10 Class III 0.00 0.00 -27.02 0.01 443.32 Price 0.78 9.06 3.78 0.00 0.00 -2.09 Revenues Total 16.69 6.74 3.88 0.00 -7.74 Class I 0.55 17.97 7.30 51.49 -0.10 -51.14 Imports Total -1.15 -0.61 -0.19 -0.11 0.01 Costs 2.45 9.92 5.86 2.35 1.95 -0.13 Regions 1 and 2 also gain. The additional cost of shipping down to region 1 causes its export price to 3ump over 9 percent. This, in turn, allows its higher cost of production industry to burgeon. Total production in region 135 1 increases over 7.3 percent with total Grade A revenues rising by approximately 16.7 percent. Overall, when trans- portation costs increase 50 percent, all other variables held constant, producers stand to come out about even (total revenues across all regions increase less than .4 percent). For consumers, the cost of increased transportation is passed directly on to them. The overall import price increases 2.45 percent. As expected, southern regions see the largest increase in import price (9.9 percent in region 1 and 5.9 percent in region 2). In contrast, the import price actually falls, though negligible, in region 9 where the decline in Class I exports floods its market. 5.7.1b Applying RO Filtration at 0.90/cwt The application of RO filtration in the face of in- creased transportation costs allows Upper Midwest producers to once again become gainers. Under transportation cost increases, the adoption of RO filtration would allow region 9 to regain its lost market and region 8 to develop new markets. Figure 5.16 illustrates the resulting Class I shipment pattern under this run (TC2RO9). 136 ’0‘ Figure 5.16. Distribution pattern under 1985 market conditions with transportation costs increased fifty percent and R0 filtration applied at 0.90/cwt, (TC2RO9) The increase in Class I sales by both region 8 and 9 are substantial, 19.5 and 101 percent respectively. However, the gains obtained through RO filtration go beyond the status quo established by the original BASE solutionl. This is illustrated by the percentage changes listed in Table 5.15. A comparison indicates that revenues in region 1 would actually fall a total of 6.45 percent versus a revenue gain of 1.4 percent in region 9. Hence, while the 1 Note that this scenario represents a simultaneous increase in transportation costs and application of R0 filtration. 137 Table 5.15. Market Impact Under 1985 Conditions with Fifty Percent Increased Transportation Costs and R0 Applied at 0.90/cwt, (TC2RO9) Regions Variable All 1 2 7 8 9 (Numbers are as a percentage change of BASE values) Exports Total -3.02 -2.92 0.00 0.00 0.32 Class I -0.12 -3.36 -3.25 -2.60 19.33 0.35 Class III 0.00 , 0.00 1.37 -2.05 0.00 Price -0.83 -3.72 -3.63 0.01 0.01 1.19 Revenues Total -6.45 I6.24 -0.19 0.19 1.41 Class I -0.94 -6.95 -6.76 -2.60 19.34 1.55 Imports Total 0.30 0.13 -0.19 -0.11 -0.24 Costs 1.44 -2.59 -1.22 2.35 1.96 3.02 introduction of RO filtration appears essential for producers in regions 8 and 9 given a 50 percent increase in transportation costs, the end result is not an even distribution of gains when R0 filtration is priced at 0.90/cwt. 5.7.2 Alternative Supply Elaaticitiaa Elaaticitiaa may be difficult to estimate even for the simplest of commodities. Judging by the range of elas- ticity sets developed and of estimation techniques emp- loyed, estimation of milk supply elasticities has not been an easy task. Accordingly, the model used in this study was tested for its sensitivity to changes in regional supply elasticities. Initially two sets of short run supply elasticities, Hallberg et al.'s and Huy’s, were considered for use in the oo< porote, genero‘ runs a elooti Hollbc icant proui initi hi 1985 the elos inpa gene in 1 reg; aUbe Tob] Undo 138 the model. In selecting which of these sets to incor- porate, base runs were generated for each. When model generated regional supply and demand levels for the two runs were compared to actual 1985 levels, Huy's initial elasticities provided slightly better results than did Hallberg at al.’s: although, the results were not signif- icantly different. In light of the base run's role of providing the opportunity for finer calibration, Huy‘s initial 1985 elasticities were selected. After all runs were generated, a revised set of Huy's 1985 elasticities became available. This increased the importance of testing the model’s reaction to alternate elasticities. As an indication of the new elasticities impact on regional supply schedules, intercept terms generated from the new and old elasticities are presented in Table 5.16. Note that while the ordinal ranking of regions remained the same, the degree of impact varied substantially. Table 5.16. Regional Supply Schedule Intercept Term Values Under New and Old Supply Elaaticitiaa Intercept Percentage Region Old New Change 139 A problem arose in generating model runs due to the negative elasticity values associated with regions 8 and 9. GTP only accepts supply elasticities which are greater than or equal to zero. To test the legitimacy of setting these to zero, region 9’s elasticity was ad3usted downward from .15 in successive runs, 915311;,pggiggg, Table 5.17 reflects how region 9’s supply schedule coefficients and the equilibrium supply quantities for all regions changed. Note that while altering region 9’s supply elasticity did significantly impact its supply schedule coefficients, there was a negligible impact on model equilibrium levels. Table 5.17. Impact of Prograaaivaly Lower Supply Schedule Elaaticitiaa on Region 9’s Supply Schedule Terms and Equilibrium Ouantitiaa Supply Elasticity Slope Intercept Region 9 All Others .15 .56 44.00 51.652 no change .10 .37 46.58 51.636 no change .05 .19 49.17 51.766 no change .00 .00 51.76 51.760 no change Given this apparent insensitivity to changes in a single region's supply elasticity, the model was rerun with region 8 and 9’s elasticities set at zero. The impact on relevant results was negligible. Table 5.18 displays the overall impact of the new elasticities on model results under 1985 market conditions (BASE). The distributional patterns and regional Class I supply and demand quantities levels remained within II- 1 percent of initial results, with the exception of regions 4 and 8. Region 4’s aupply increased 140 2.5 percent while region 8’s increased 1.8 percent. Interestingly though, all of this additional supply was shipped to region 10, the Class III sink. Hence, the effect of the revised elasticities was insignificant in relation to the analysis presented in this chapter. Table 5.18. Regional Impact of "New“ Supply Elaaticitiaa on Equilibrium Ouantity and Price Levels Region Supply Demand Price Old New Change Old New Change Old New (mil. cwt) (8) (mil. cwt) (X) (0/cwt) 1 20.2 20.2 0.0 23.1 23.1 0.0 17.08 17.05 2 27.8 27.9 0.0 34.4 34.4 0.0 15.72 15.69 3 64.6 64.5 0.0 58.1 58.2 0.0 ' 14.86 14.94 4 66.4 68.1 I2.5 50.5 50.5 0.0 13.31 13.28 5 53.0 52.9 -0.3 44.1 44.1 0.0 14.83 14.81 6 65.4 65.9 I0.7 44.5 44.5 0.0 14.06 14.06 7 241.8 243.2 I0.8 118.1 118.1 0.0 14.57 14.57 8 242.8 247.2 +1.8 23.3 23.3 0.0 12.98 12.98 9 51.6 51.8 I0.4 23.1 23.1 0.0 13.67 13.67 10 412.9 421.4 I2.1 Although there is little doubt that drastic alterations to supply and demand elasticities would impact model results, the test conducted in this section indicates that the model is reasonably insensitive to moderate elasticity changes. This is especially true for near proportional changes. Hence, the results and analysis presented in this study, developed using Huy’s initial 1985 elasticities, provide an acceptable level of accuracy. 5.8 Comparison of Results with Other Studies This section addresses the cross comparison of results presented in this thesis with those from other dairy node: Nova] Whip] Altht this gene regi onon broo diff Thei Pric Prod enti cost filt With or“ “lac it c 0r . Furt 141 modeling studies. Specific studies addressed are those by Novakovic at al., DAMPS, (1980): Hallberg et al. (1978): Whipple (1983): and, Hammond, Buxton and Thraen (1979). Although these studies most closely follow the design of this thesis, they differ in specification and intent. In general, the time period covered, data incorporated, regionalization, and issues addressed vary considerably among studies. Given these intrinsic differences, only broad comparisons are made and general insights gained. The DAMPS study addressed the realigning of regional differentials according to actual transportation costs. Their results indicate that in no region does the resultant pricing structure lead to a near or total reduction in production. In contrast, within this thesis the differ- entials were held constant and, when the transportation cost structure changed significantly with adoption of RO filtration, several production regions experienced a total withdrawal from the market. Although the two alternative approaches to handling Class I differentials do prohibit direct comparison, they also help to illustrate one effect of the short-run model: it does not allow for tho industry's ability to make policy or marketing ad3ustments in response to a new technology. Furthermore, there are several areas where general DAMP’s results coincide with those obtained within this study. For example, producer price and marketings in the Northeast free enti cope they Pro; enti in C 99ne: exPer l‘enov ‘nd 1: 13.7 p upPTOx 142 are found to be highest when the Class I price and differ- ential are highest in that region: consumption varies relatively minimally with changes in the Class I price: producers in the Upper Midwest do better the higher the Class III price: and the Southeast and Southcentral regions’ prices are lowest when Class I differentials are set according to the cost of shipping milk in ingredient form. While these observations correspond with the results obtained within this thesis, they are more comparisons with accepted beliefs and theory than a good test of model results and they do not provide any new understanding of spatial market response under RO filtration. Hallberg at al. focused on equalizing the prices of fresh and reconstituted milk via altering Class I differ- ential levels. As with the DAMPS study, there is limited capacity for cross comparisons to be made. For example, they found that consumer expenditures increase by a greater proportion than producar revenues when Class I diffar- entials are increased. Although no proportional increases in Class I differentials are made within this thesis, the general relationship between producer revenues and consumer expenditures is exhibited when Class I differentials are removed entirely: Grade A export revenues fall between 7.5 and 11.5 percent while Class I import expenditures fall by 13.7 percent. Additionally, Hammond at al. note that approximately 60 percent of shipments to the Northeast :eglOT pondi of i‘ non to Sta reg pro ohc Pri R0 of 143 region’s demand centers are intraregional. The corres- ponding region in this study, region 7, supplied 70 percent of its Class I demand. Whipple found that if ingredients were priced at their manufactured price then farm level prices would fall by up to 15 percent in Florida and by up to 2 percent in the Lake States. Furthermore, the gross farm receipts across all regions would drop. Although this thesis expresses producer prices as export prices, responds only in the short-run, and does not directly deal with the issue of pricing reconstituted milk (due to only Grade A milk being RO filtrated), it does produce similar responses in terms of direction of impact. For example, when RO is priced at 0.30/cwt (yielding a transportation cost function nearest to that for shipping milk in ingredient form), export prices in Florida fall by over 13 percent, export prices in the Upper Midwest increase by approximately .25 percent, and total export revenue for Grade A milk drops between 5 to 40 percent across all regions (depending upon the actual anount of Grade A produced for Class III use). Again these results suggest compatibility between research results in very general terms. Hammond, Buxton, and Thraen approached analyzing the regional impact of reconstituted milk on regional Class I differentials and production via altering the differentials and The to noi ch oi IE 144 and via pricing ingredients at their manufactured price. The first method is not used in this study and the second is not applicable: thus, direct comparison of results is not feasible. 0f general interest however, they find that changing supply elasticities proportionally has a rather minor influence on results. While this is analogous to results presented in this chapter, it is more an indication of model sensitivity to elasticity changes than it is to the markets’ reaction to a policy or technology change. 5.9 Summary In this chapter the model specified in Chapter Four was directly applied to the dairy industry under several different policy and economic scenarios. When tested against actual 1985 marketings, the model responded with reasonable accuracy. As more drastic policy changes were implemented and the model was stretched further beyond its original design, less confidence was held in any one result. However, even under these more extreme conditions, the model does provide insight as to what the market impact of alternative scenarios would be under a perfectly competitive, spatial equilibrium framework. Throughout all model runs, changes to regional Class I sales, revenues, and distribution patterns were used as indications of the possible market impact of alternative scenarios. Of specific interest was the viability of R0 filtration within the model. 145 Although results under alternative scenarios varied significantly, across all model runs it was found that the introduction of R0 filtration would benefit importers at the expense of exporters. Regionally, the impact was inconsistent across runs. In general, exporters in the lower relative cost of production and Class I differential regions saw prices and revenues fall less than did ex- porters in other regions. The converse was true for importers. In terms of impact on the two low cost of production regions, 8 and 9, introduction of RO filtration had a mixed impact under alternative scenarios. Under the BASE scenario series, only region 8 gained export revenues under RO filtration priced at 0.30 and 0.90/cwt; region 9 saw no revenue change under R0 at any price. During SEPT condi- tions, adoption of RO filtration actually reduced region 8’s revenues while increasing region 9’s. When Class I differentials are removed, R0 filtration had no significant impact on any region under any pricing scheme. It appears region 9 benefits substantially from full scale adoption of R0 filtration under the two remaining scenarios. First, under the new 1986 differentials both regions 8 and 9 experienced substantial revenue gains via RO filtration. Second, under a significant (50 percent) increase in transportation costs and R0 filtration priced at either 0.30 or 0.90/cwt, region 9 en30yed substantial revenue gains: region 8 experienced no change in revenues. CHAPTER SIX SUHHARY AND CONCLUSIONS This study examined the potential economic feasibility and market impact of applying reverse osmosis filtration to fluid milk. The motivation for this research stemmed from a combination of factors. Primary among them were the tightening of financial resources within the industry and improvements in reverse osmosis technology. A short-run spatial equilibrium model of a selected segment of the U.S. Grade A milk market was developed and applied under a range of pricing and policy scenarios. Specific questions asked were: (1) Who stands to gain from the nation-wide adoption of RO filtration? (2) How will production shift and what are the regional implications of such a change? And (3) Is RO filtration even potentially feasible in the political economic marketing sense? This thesis has approached these questions through both descriptive and quantitative analysis. Ragional compet- itive advantage in milk production was established. The key marketing variables affecting and potentially affected by the industry-wide adoption of RO filtration for Class I use were identified. The current policy constraints to 146 147 marketing reconstituted milk were discussed. A brief summary of the motivation behind this research, the model utilized, and the results obtained are presented below. The remainder of the chapter focuses on the implications of model results and on conclusions which can be drawn from them. 6.1 Background to Research Issue The U.S. dairy industry is immersed in an era of tran- sition. In an effort to decrease government expenditures and reduce the milk surplus, the dairy support price has been lowered by over two dollars per hundredweight in the past five years. Under the burden of heavy debt and reductions in the support price, many dairy operations have been forced out of production. In contrast to these pressures to reduce supply, productivity expanding tech- nologies, such as bovine growth hormones, are on the horizon. The result of these opposing forces is both increased tension and an intensified effort to identify means by which to minimize the negative effects of tran- sition and achieve greater economic efficiency. Technological advances related to the long distance transportation of bulk milk have made the topic of greater efficiency gains through comparative advantage relevant. Of primary interest is the advance in reverse osmosis (R0) filtration. This process would allow bulk milk to be reduced in volume and weight by fifty percent and would 148 provide a fluid product meeting taste, consistency and nutritional requirements of the consumer. While the present regulatory environment restricts the economic marketing of an R0 filtrated milk product, given a positive environment, RO filtration could be an attractive means by which transportation costs could be reduced and efficiency increased. In the process, the dairy industry could be aligned according to competitive advantages in production. Such a realignment would allow the industry to achieve increased economic efficiency. It appears this may be of importance to the industry’s long-run success. 6.2 The Model An annual spatial equilibrium model was specified according to the market characteristics of the dairy industry and the ob3ectives of this study. The modeled area was broken down into nine supply and ten demand regions covering 29 federal orders and 33 states. Esti- mated costs of applying RO filtration, transportation cost functions, supply and demand elasticities, and regional data were all incorporated into the model. The fully specified model was then applied to estab- lishing the potential economic feasibility and regional and market impact of RO filtration on the Class I milk market. To capture the range of possible impacts, several pricing and policy scenarios were developed. Specific scenarios were as follows: (1) three possible costs of applying R0 149 filtration were incorporated into the base model; (2) the model was ad3usted to reflect market conditions during September, when supply is low, demand is high and inter- regional shipments are greatest: (3) a significant cut in CCC purchaaea precipitating a drop in the support price was simulated: (4) the total removal and the realignment of Class I differentials were incorporated: (5) the impact of a fifty percent increase in transportation costs was simulated: and finally, (6) alternative supply elasticities were applied to determine the model’s sensitivity to that important parameter. Solutions were then generated through the Generalized Transportation Problem microcomputer program developed by Holland and Sharplea (1984) and Holland (1985). Results from each of these alternative market scenarios were compared to their respective baseline runs on a percentage change basis. Under each scenario the primary questions addressed were (1) how supply and demand quan- tities were affected, (2) how the distributional pattern changed, and (3) what was the resulting impact on costs and revenues. Specific interest was paid to the regions believed to experience the greatest impact under scenario changea: Florida, Southeast, Northeast, Minnesota-Wiscon- sin, and_Michigan. 150 6.3 Summary of Results Model generated results were both insightful and, in some cases, surprising. The fully specified model gen- erated baseline results reflecting 1985 market levels with acceptable accuracy. This provided the foundation upon which alternative scenario results were compared in order to determine the potential market impact of policy and price changes. 6.3.1 R0 Filtration When R0 filtration was applied to the baseline model, regional impacts were significant. Distributional changes tended to favor the Minnesota-Wisconsin region to the detriment of the Florida, Southeast and Northeast regions. In terms of producer revenues, the overall impact of RO filtration was a significant loss in total revenues. Regionally, producers in Florida, the Southeast and Northeast suffered the greatest loss in revenues. Under each pricing scheme, the benefits of full scale adoption of RO filtration under 1985 market conditions were passed directly on to consumers. The regions which saw the greatest fall in producer prices also saw the largest fall in consumer prices. 6.3.2 September The impact of R0 filtration under September market conditions remained substantial. The Minnesota-Wisconsin 151 region gained markets while the two southernmost regions sustained the largest declines in price. The net impact to all producers continued to be negative, however, the degree of this impact was less than experienced under the baseline 1985 scenario. Regionally, consumer and producer prices once again moved together. 6.3.3 Reduction in C00 Purchases When CCC purchases were limited to 5 billion pounds, the M-W price dropped to 010.32. This, in turn, led to a significant decline in the weighted average Grade A market price. The inelastic nature of milk demand limited the impact of this price fall to a relatively minor change in demand: hence, the ma3ority of the 8.2 billion pound decrease was acheived through reduced supply. Producers in the relatively high cost of production regions absorbed the greatest impact of the reduction both in terms of fallen prices and sales. The Northeast region was the only region to experience a significant increase in Class I sales but no region saw an increase in total revenues. The most significant revenue loss hit the Florida, Southeastern, Northeastern and Minnesota-Wisconsin regions. When RO filtration was applied under this scenario, regions with relatively low Class I differentials bene- fited. As a result, many of the regions which suffered the largest losses under decreased CCC purchases received the greatest benefit from the adoption of R0 filtration. 152 Specifically, model results indicate that producers in the Minnesota-Wisconsin and Michigan regions benefited signifi- cantly by the market-wide adoption of R0 filtration in the face of significant reductions in CCC purchases or the M-W price. Producers in the Florida, Southeast and Northeast regions, however, experienced compound negative impacts. 6.3.4 Change in Class I Differentials Model results under a one hundred percent removal of Class I differentials indicate that interregional shipments of Class I milk decrease and the market wide Class I price falls substantially. Class I revenues drop dramatically but surprisingly, the greatest loss is felt by the Michigan region. In contrast, the Florida region becomes an isolated market, experiencing a negligible fall in Class I price and an increase in total revenues. It appears from this that Class I differentials actually serve to subsidize exports to certain regions -- perhaps preventing local producers from supplying local Class I markets. The application of RO filtration led to a reduction in prices but no alteration in distributional patterns, as would be expected. When 1986 differentials were substituted, revenues and export prices remained unchanged overall, but regionally, shifts occurred. The Michigan region received the greatest gain in overall Grade A price while the Florida region received the greatest loss. Regional revenue gains favored. 153 the two Upper Mid-west regions as the "export subsidy" characteristic of relative Class I differentials was enhanced. When RO filtration was implemented, the imbalance of benefits swung even wider. The Minnesota-Wisconsin region’s Class I sales increased dramatically to the detriment of the two southernmost regions, the Southern Plains and the Northeast. 6.3.5 Increased Transportation Costs A fifty percent increase in transportation costs had the effect of altering the impact of Class I differentials. As a result, a decrease in interregional Class I shipments and an increase in intraregional sales occurred. The Florida region actually became self-sufficient while other regions imported their additional supply from closer markets. Hence, the ability of the low cost of production Upper Midwest regions to capitalize on their competitive advan- tage was limited. Minnesota-Wisconsin producers were most adversely affected while their counterparts in the Florida, Southeast and Northeast regions experienced a significant increase in total revenues. Consumers paid directly for the increased cost of trans- portation: likewise, when RO filtration was applied they benefited directly from the reduction in transportation costs. Producer benefits from the application of RO filtration were primarily allocated to the Michigan and 154 Minnesota-Wisconsin regions. This suggests that, in the face of increasing transportation costs, all other factors held equal: in order for producers in those regions to maintain revenues, they may find it necessary to adopt a bulk reduction technology such as R0 filtration. 6.4 Points of Caution The model utilized to gain these insights is simplistic relative to the dairy industry itself. It can only give the theoretical solution to the spatial equilibrium model as specified. It is a short-run annual model and as such is not capable of capturing all the subtleties of the market’s long-run ad3ustment to a new technology. The model produces an instantaneous shift to a new equilibrium position while in reality the market ad3usts in a far more interactive and dynamic manner. For example, it is inevitable that policy changes would be made during tran- sition, altering the path towards the new long-run equilib- rium. Given this understanding, it would be misleading to extract any given model generated coefficient and compare it to the true value found in an isolated, dynamic Grade A market. Even with these caveats, aignificant insights were gained in approaching the general ob3ectives of this research. 155 6.5 Conclusions The impact of RO filtration on the Grade A milk market, as indicated by model solutions, could be significant. However, the overall impact appears to effect producers adversely while providing significant benefits to con- sumers. This remains true across all pricing and policy scenarios. Economic theory suggests that RO filtration would be implemented if sufficient gains could be captured to cover all costs. Therefore, it would be necessary for consumers to properly compensate producers and for the underlying political costs to remain unprohibitive. The model can not account for these elements. In terms of industry dynamics, market power and bar- gaining among producer cooperatives play a very real and important role. The industry does not operate according to perfectly competitive market theory. Recognizing this, it is reasonable to suggest that some of the benefit gained by consumers in the theoretical case could be usurped by producers. Across scenarios and regions this represents a possible average revenue gain of four percent. In an industry where profit margins are slim and producers are struggling to remain viable, this represents a significant increase. How much of that gain could be negotiated away from consumers and how it would be distributed regionally is unclear and not the sub3ect of this research. One can merely say that under the model developed and presented 156 within this thesis, it is not likely that R0 filtration would find use within the U.S. Grade A milk market at this time. Actual long-run results would likely differ from those presented here to some unknown degree. To summarize the implication of model results, the full scale adoption of R0 filtration within the Grade A milk market would depend heavily on three related factors. First, the magnitude of the potential gain to consumers and the degree to which that gain could be transferred to producers must be large enough to instigate the necessary changes. Second, the gain captured by producers must be large enough to cover the indirect costs of policy changes (i.e. lobbying and developing new market institutions) and it must be large enough to compensate the regional losers. Third, the adoption of R0 filtration would be dependent upon development of a set of mechanisms facilitating these compensatory transfers between producers. This thesis neither addresses questions related to the social welfare and policy implications of R0 filtration nor does it confront issues of market efficiency and equity. These are important areas to be researched when adoption of R0 filtration is under serious consideration. This thesis merely develops and presents a model with the ob3ectiva of testing that model and providing insights as to possible changes within the industry. The results obtained together with the analysis presented meet that ob3ectiva. GLOSSARY OF TERMS GLOSSARY OF TERMS Allocation Provision: The FMMO accounting system whereby either Class I milk not originating from a federal order or reconstituted milk is matched in volume by ”up- allocated", non-Class I milk. Such milk in excess of what is available to up-allocate is charged a Compensa- tory Payment equal to the local Class I differential. Together, allocation provisions and compensatory pay- ments ensure that the cost of non federal order or reconstituted milk is at least as great as the minimum Class I price within a given federal order; thereby encouraging use of local supplies and effectively removing the incentive to reconstitution. Blended Milk Product: A product made from a blend of fresh whole milk, water and nonfat milk powder or butter oil. Blended milk allows an area’s milk supply to be stretched. Blend Price: The price which producers receive for their Grade A milk pooled within federal orders. The blend price is a weighted average price determined according to the proportion of milk allocated to the alternative Grade A classes. Call Provision: A FMMO provision implemented at the discretion of the Market Administrator and which forces a set amount of milk to be channeled into fluid use, in times of deficient local supply. Capper-Volstead Act: A 1922 Federal act allowing producers to organize for the purpose of buying and selling farm products cooperatively. In the absence this legislation, producer cooperatives were sub3ect to antitrust suits. Casein: A protein component found naturally in milk and cheese. During the R0 filtration process, casein often becomes clogged on the membrane surface, reducing the the system’s effectiveness and requiring thorough cleaning of the membrane. 157 158 Classified Pricing: The pricing system under which pro- cessors regulated by federal order provisions pay for Grade A milk according to the class in which it is used. Class I Hilk: Grade A milk sold for fluid consumption in federal milk marketing orders. Class II Nilk: Grade A milk sold for use in soft manufac- tured products, such as sour cream, yogurt and cottage cheese, under a FHHO with three Grade A classes. Where only two classes exist, Class II comprises Grade A milk used in any form of manufacturing. Class III Hilk: Grade A milk sold in federal orders with three classes and used in the manufacturing of hard products, such as cheese, butter and milk powder. Class I Differentials: The assessment added onto the M-w price for Grade A milk sold in federal orders and going to Class I use. Class I differentials from a price surface which increases with distance from Eau Claire Wisconsin, the base pricing point. Commodity Credit Corporation (CCC): Government operated organization through which storeable dairy products are purchased at a set price. CCC purchases directly serve to support the market price of manufactured dairy products and indirectly support the price of all milk. Compensatory Payments: A FMMO surcharge assessed on either Class I milk not originating from a federal order or reconstituted milk above what has been up-allocated (see Allocation Provision). The payment is equal to the local Class I differential. Flux: The rate at which the solution passes through the membrane. The flux within an R0 system is a function of total solids in solution and the pressure under which the solution flows through the system. Formula Pricing: An institutional pricing system whereby a given commodity’s price is calculated from a formula incorporating economic variables related to the commodity’s value and which act as "price movers". Examples of such variables are cost indices and prices of substitute and complementary goods. Grade A Milk: Milk meeting fluid comsumption health and quality standards. Only Grade A milk is regulated under federal orders. Grade 3 Milk: Milk meeting health and quality standards for manufactured use but not for fluid use. 159 H-U Prica Series: The USDA estimated average price paid per hundredweight of milk, f.o.b., by manufacturing plants in the states of Hinnesota and Wisconsin. Osmosis: The naturally occurring process in all organisms where water and a solution are separated by a membrane. Under osmosis, the water passes through the membrane to dilute the solution until the pressure exerted on both sides of the membrane are equal. Osmotic Pressure: The force exerted by a solution against the membrane system. Each solution exerts a different osmotic pressure. Over-Order Premium: An additional charge negotiated by producer cooperatives and paid by handlers for milk going towards Class I use. This premium is often associated with the costs of marketing services such as transportation, full-supply agreements, and handling of Class III milk. Permeate: The liquid which has passed through the membrane within the reverse osmosis process. Phase Change: A phase change occurs when a discrete homogeneous characteristic of the solution is separated from the rest of the solution by some external force. For example, a phase change can be evidenced by altered taste, consistency, color or odor of the solution. Price Linkage: The collection of marketing variables which act as a link between regional equilibrium price levels. Examples of these linkage variables are transportation costs, over-order premiums, and Class I differentials between regions. Retantate: The substance remaining after a removal process, such as R0 filtration, is complete. Reverse Osmosis: When sufficient pressure is applied to the solution to offset its osmotic pressure. Under reverse osmosis, the water existing within the solution is forced through the membrane leaving a more concen- trated solution behind. Reverse Osmosis (R0) Filtration: A pressure driven mem- brane separation technique applied to liquid substances such as milk and cheese whey. The primary difference between R0 filtration and ultrafiltration is the finer degree of particle separation obtained with R0. 160 Spatial Equilibrium: The achievement of an equilibrium price surface, across regions, at which regional equilibrium price levels differ by the price linkage and the total quantity produced across all markets is exactly equal to the total quantity demanded. Ultrafiltration (UP): A pressure driven membrane aeparation technique applied to liquids, such as milk, for water removal. whey: The part of milk remaining after the cheese making process. APPENDICIES APPENDIX A BUPLEIEITAL EOUATIOIS TO CHAPTER THO Tha markat variables which exist within the dairy industry, as outlined in Chapter Two, are complex and difficult to model in general terms. The pricing mechan- isms and ralationships which where discussed can be placed within tha general fluid milk marketing model as follows. I Processors Price: The Class I price in any region i is: pI1 . pH-U1 e 01 e 01 where, P11 I the Class I price P1111 I the Hanufactured goods price (lagged two months) D; I the Class I differential 01 I the region's average over-order premium The Class III price for any region i is approximately: pIIIt . pH-w e v where V I the weighted change in gross value of milk used to make cheddar cheese and butter/nonfat dry milk. 161 162 I Producer price: For each region i the blend price is: p81 . (pI£QI1 e pIIItoIIIi)/(in + 01111) where: 011 l (011 + 01111) I Class I utiliza- tion ratio, U1, in region is thus, 981 I P11DU11 + pIIItaullli for .ny region 1. I Demand Functions Demand for Class I use: D1 I a + bPI Demand for Class III use: D111 I c + dPIII Demand for Grade B milk: Db I e * be I Supply Functions Supply of Grade A milk: 8‘ I a + bPB Supply of Grade 8 milk: Sb I c + de I Equilibrium occurs where total supply equals total demand across all regions. APPENDIX D SUPPLEIEHTAL EOUATIOIS TO CHAPTER TWO Table 3.1. Regionally Adjusted 1985 Annual Average Supply, Demand, Price and Class I Differential Levels Quantity Quantity Class I Differentials Region Supplied Demanded Price 1985 1986 (mil. cwt) (dollars per cwt.) 1 2,021 2,310 14.09 3.03 4.27 2 2,785 3,454 13.30 2.48 3.30 3 6,369 5,854 12.48 2.20 3.07 4 6,890 5,046 12.78 1.44 1.69 5 5,259 4,438 12.80 1.94 2.52 6 6,618 4,550 12.70 1.68 2.03 7 24,376 11,843 12.00 2.79 3.05 8 24,720 2,323 12.60 1.20 1.31 9 5,176 2,318 12.28 1.60 1.75 10 na 42,079 na na na Table 3.2. Regionally Adjusted Demand, Supply and Revised Supply Elaaticitiaa Revised Region Supply Demand Supply 1 .811 -.113 .500 2 .805 -.100 .496 3 1.113 -.091 .775 ‘ e 567 " e 078 a 183 5 .793 -.089 .468 6 .584 -.078 .172 7 .680 -.078 .133 8 .300 -.057 0 9 .268 -.078 0 163 164 Table 8.3. Regional Weighted Average Class I Utilzation, Population and Per Capita Consumption, 1985 Adjusted Class I Per Capita Region Utilization Population Consumption 1 88 11,215 206 2 79 15,923 217 3 64 26,345 222 4 45 19,517 258 5 71 20,463 217 6 58 18,234 250 7 45 55,085 215 8 17 8,966 259 9 42 9,100 255 Table 8.4. Hileage Matrix Between all Supply and Demand Regions, One-Way Niles Supply Demand 1 2 3 4 5 6 7 8 9 1 30 488 1,200 1,322 840 1,058 1,261 1,486 1,229 385 86 892 932 463 732 985 1,090 838 1,048 738 93 917 1,010 1,122 1,628 1,250 1,181 1,289 829 762 122 604 519 986 335 280 684 377 1,146 1,042 210 898 564 948 624 1,018 655 1,004 609 355 50 584 698 276 1,168 913 1,491 1,108 708 504 128 1,081 613 1,498 1,045 1,067 387 811 622 1,153 0 505 00‘100‘ (0” 1,256 813 1,158 644 551 223 760 572 55 APPENDIX C NODEL GENERATED SOLUTIONS: ALL REGIONS ALL RUNS Run Title Description BASE BROS BRO9 BRO175 CCC CCCRO9 ND NDRO9 NDCCC NDCCRO9 D86 086RO9 Initial annual run serving as the standard for comparisons. 1985 market conditions with no R0 applied. BASE specification with R0 incorporated at a cost of 8.30/cwt. BASE specification with R0 incorporated at a cost of 8.90/cwt. BASE specification with R0 incorporated at a cost of 81.75/cwt. BASE model ad3usted for reduced CCC purchases. An import quote was placed on Class III milk, reducing purchased by 444 million pounds. CCC specification with R0 incorporated at a cost of 8.90/cwt. BASE model ad3usted for the full removal of Class I differentials. No R0 applied. ND specification with R0 incorporated at a cost of 8.90/cwt ND specification with reduced CCC purchases ND specification with reduced CCC purchases and R0 incorporated at 0.90/cwt BASE model with 1986 Class I differentials substituted for the 1985 levels. No R0 applied. D86 specification with R0 incorporated at a cost of 8.90/cwt. 165 166 Figure 0.1. (can't) TC2 TC2RO9 SEPT SR03 SR09 SROl75 HUYOO HUY05 HUYlO HUY15 BASE model with a fifty percent increase in transportation costs incorporated. No R0 applied. TC2 specification with R0 incorporated at a cost of8.90/cwt. nodal run generated based on September 1985 market conditions. Serves as a base upon which R0 feasibility during month when shipments are high. No R0 applied. SEPT specification with R0 incorporated at a cost of 8.30/cwt. SEPT specification with R0 incorporated at a cost of 8.90/cwt. SEPT specification with R0 incorporated at a cost of 01.75/cwt. Base model with Huy’s ad3usted supply elasticities incorporated (region 8 and 9’s elasticities set equal to zero) HUY specification with region 9’s supply elasticity sat at .05 "UV specification with region 9's supply elasticity sat at .10 HUY specification with region 9’s supply elasticity sat at .15 Figure C.1. Reference to model generated solution titles and discriptions Figure C.2. Delineation of model regions 167 FNNO’s Region Encompassad States Encompassad Demand Center Supply Center 1 Upper Florida Tampa Bay SE Florida Florida Lake Wales,FL Lakeland,FL 2 AL-Ueat FL Georgia New Orleans-NS Nacon,GA Georgia Mississippi South Carolina 3 Central AR Greater LA Lubbock TX Panhandle SW Plains Texas Arkansas Louisiana Oklahoma Texas Lufkin,TX Greenville,TX 4 Central IL Iowa Southern IL St.Louis-Ozarks Illinois Iowa Nissouri Galesberg,IL 5 Louis-Lex-Evan Nemphis Nashville Puducah TN Valley Kentucky North Carolina Tennessee Virginia Nount Airy,NC 6 Indiana Ohio Valley Indiana Ohio west Virginia Newark,0N 7 Nid Atlantic New England NY-NJ E.ON-U.PA Connecticut Delaware Naine Naryland Nassachusetts New Hampshire New Jersey New York Pennsylvania Rhode Island Vermont Port Jarvis,NY 0neonta,NY 8 Chicago Reg Upper lid-West Ninnesota Wisconsin 10 All of above Figure 0.3. All of above All of above Federal milk marketing order (FNNO’s) and states encompassed within regions, regional demand and regional supply centers 168 Table 0.1. Import/Export Ouantitiaa and Prices for All Regions Under 1985 Narkat Data (BASE) Exports Imports Region Importer Ouantity Price Exporter Ouantity Price (mil.cwt.)(0/cwt) (mil.cwt.)(G/cwt) 1 1 18.18 17.08 1 18.18 17.44 10 2.02 11.78 2 3.34 17.44 Totall 20.20 Totall 23.05 2 1 3.35 15.72 2 21.68 16.46 2 21.68 15.72 4 8.99 16.46 10 2.79 11.78 5 3.72 16.46 Total 27.81 Total 34.38 3 3 58.19 14.86 3 58.19 15.65 10 3.67 11.78 Total 58.19 Total 64.56 4 2 8.99 13.31 4 50.52 14.00 4 50.52 13.31 Total 50.52 10 6.89 11.78 Total 66.40 5 2 3.72 14.83 5 44.05 15.91 5 44.05 14.83 Total 44.05 10 5.26 11.78 Total 53.03 6 1 1.52 14.06 6 45.45 14.56 6 45.45 14.06 Total 45.45 7 11.83 14.06 10 6.62 11.78 Total 65.42 7 7 83.05 14.57 6 11.83 15.29 10 158.27 11.78 7 83.05 15.29 Total 241.32 9 23.23 15.29 Total 118.12 8 8 23.31 12.98 8 23.31 12.99 10 219.50 11.78 Total 23.31 Total 242.81 9 7 23.23 13.67 9 23.14 14.20 9 23.14 13.67 Total 23.14 10 5.18 11.78 Total 51.55 1 Totals may not equal the sum of the parts due to rounding 169 Table 0.2. Import/Export Ouantitiaa and Prices for All Regions Under 1985 Narkat Data with R0 Applied at 8.30/cwt (BROS) Exports Imports Region Importer Ouantity Price Exporter Ouantity Price (mil.cwt)(s/cwt) (mil.cwt.)(s/cwt) 1 10 18.03 11.78 8 23.46 14.75 Total1 13.03 Total1 23.46 2 10 25.74 11.78 8 34.83 14.44 Total 25.74 Total 34.83 3 3 54.25 14.04 3 54.25 14.83 10 6.37 11.78 4 4.24 14.83 Total 60.62 Total 58.49 4 3 .34 13.23 8 50.56 13.85 10 61.94 11.78 Total 50.56 Total 66.17 5 5 44.41 13.73 5 44.41 14.56 10 5.47 11.78 Total 44.41 Total 49.88 6 6 45.52 13.75 6 45.52 14.25 7 12.46 46.75 Total 45.52 10 6.62 11.78 Total 64.60 7 10 241.37 11.78 6 12.46 14.11 Total 241.37 8 83.24 14.11 9 23.15 14.11 Total 118.85 8 1 23.46 12.99 8 23.31 12.99 2 34.83 12.99 Total 23.31 4 50.56 12.99 7 83.24 12.99 8 23.31 12.99 10 27.44 11.78 Total 242.83 9 7 23.15 13.60 9 23.15 14.13 9 23.15 13.60 Total 23.15 10 5.18 11.78 Total 51.48 1 Totals may not equal the sum of the parts due to rounding 170 Table 0.3. Import/Export Ouantitiaa and Prices for All Regions Under 1985 Narkat Data with R0 Applied at 0.90/cwt (BR09) Exports Imports Region Importer Ouantity Price Exporter Ouantity Price (mil.cwt)(0/cwt) (mil.cwt.)(8/cwt) 1 1 16.17 14.98 1 16.17 15.35 10 2.02 11.78 8 7.20 15.35 Total1 13.19 101.11 23.37 2 2 23.00 14.30 2 23.01 15.04 10 2.79 11.78 8 11.69 15.04 Total 25.79 Total 34.69 3 3 57.14 14.64 3 57.14 15.42 10 6.37 11.78 4 1.13 15.42 Total 63.51 Total 58.27 4 3 1.13 13.23 4 50.54 13.92 4 50.54 13.23 Total 50.54 10 14.50 11.78 Total 66.17 5 5 44.35 13.73 5 44.35 14.80 10 5.53 11.78 Total 44.35 Total 49.88 6 6 45.52 13.75 6 45.52 14.25 7 12.45 13.75 Total 45.52 10 6.62 11.78 Total 64.59 7 10 241.35 11.78 6 12.45 14.71 Total 241.35 8 82.87 14.71 9 23.15 14.71 Total 118.48 8 1 7.20 12.99 8 23.31 12.99 2 11.69 12.99 Total 23.31 7 82.87 12.99 8 23.31 12.99 10 117.76 11.78 Total 242.82 9 7 23.15 13.60 9 23.15 14.13 9 12.15 13.60 Total 23.15 10 5.18 11.78 . Total 51.48 1 Totals may not equal the sum of the parts due to rounding 171 Table 0.4. Import/Export Ouantitiaa and Prices for All Regions Under 1985 Narkat Data with R0 Filtration Applied at 81.75/cwt (BR0175) Exports Imports Region Importer Ouantity Price Exporter Quantity Price (mil.cwt)(l/cwt) (mil.cwt.)(8/cwt) 1 1 16.98 15.83 1 16.98 16.20 10 2.02 11.78 8 6.26 16.20 Total1 13.00 101.11 23.24 2 2 24.21 15.15 2 24.21 15.89 10 2.79 11.78 5 1.95 15.89 Total 27.00 8 8.35 15.89 Total 34.51 3 3 58.19 14.86 3 58.19 15.65 10 6.37 11.78 Total 58.19 Total 64.56 4 4 50.54 13.22 4 50.54 13.92 10 15.62 11.78 Total 50.54 Total 66.16 5 2 1.95 14.27 5 44.20 15.38 5 44.20 14.27 Total 44.20 10 5.26 11.78 Total 51.41 6 6 45.45 14.06 6 45.45 14.56 7 13.36 14.06 Total 45.45 10 6.62 11.78 Total 65.42 7 7 81.52 14.57 6 13.36 15.29 10 159.81 11.78 7 81.52 15.29 Total 241.33 9 23.23 15.29 Total 118.12 8 1 6.26 12.98 8 23.31 12.99 2 8.35 12.98 Total 23.31 8 23.31 12.98 10 204.91 11.78 Total 242.81 9 7 23.23 13.67 9 23.14 14.20 9 23.14 13.67 Total 23.14 10 5.18 11.78 Total 51.55 1 Totals may not equal the sum of the parts due to rounding 172 Table 0.5. Import/Export Ouantitiaa and Prices for All Regions Under 1985 Narkat Data with 000 Purchases Reduced and no R0 Applied (CCC) Exports Imports Region Importer Ouantity Price Exporter Ouantity Price (mil.cwt)(s/cwt) (mil.cwt.)(0/cwt) 1 1 16.78 15.62 1 16.78 15.99 10 2.02 10.32 6 6.49 15.99 Total1 10.30 Total1 23.27 2 2 23.11 14.38 2 23.11 15.12 10 2.79 10.32 4 3.28 15.12 Total 25.89 9 8.29 15.12 Total 34.68 3 3 56.67 14.54 3 ’56.67 15.33 10 6.37 10.32 4 1.64 15.33 Total 63.04 Total 58.30 4 2 3.28 11.96 4 50.89 12.66 3 1.64 11.96 Total 50.89 4 50.89 11.96 10 6.89 10.32 Total 62.70 5 5 44.37 13.64 5 44.37 14.71 10 5.26 10.32 Total 44.37 Total 49.63 6 1 6.49 12.60 6 45.81 13.10 6 45.81 12.60 Total 45.81 7 2.58 12.60 10 6.62 10.32 Total 61.49 7 7 103.15 13.12 6 2.58 13.83 10 121.84 10.32 7 103.15 13.83 Total 224.99 9 13.30 13.83 Total 119.03 8 8 23.45 11.53 8 23.45 11.53 10 211.53 10.32 Total 23.45 Total 234.98 9 2 8.29 12.21 9 23.33 12.74 7 13.30 12.21 Total 23.33 9 23.33 12.21 10 5.18 10.32 Total 50.09 1 Totals may not equal the sum of the parts due to rounding 173 Table 0.6. Import/Export Ouantitiaa and Prices for All Regions Under 1985 Narkat Data with Reduced CCC Purchases and R0 Filtration Applied at 8.90/cwt (CCCRO9) Exports Imports Region Importer Ouantity Price Exporter Ouantity Price (mil.cwt)(s/cwt) (mil.cwt.)(8/cwt) 1 1 15.20 13.97 1 15.20 14.34 10 2.02 10.32 8 8.32 14.34 101.11 17.22 Total1 23.52 2 2 21.57 13.29 2 21.57 14.03 10 2.79 10.32 8 13.35 14.03 Total 24.35 Total 34.92 3 3 52.67 13.72 3 52.67 14.50 10 6.37 10.32 4 5.93 14.50 Total 59.04 10.32 Total 58.60 4 3 5.93 12.30 4 50.80 12.99 4 50.80 12.30 Total 50.80 10 6.89 10.32 Total 63.62 5 5 43.11 13.19 5 43.11 14.27 10 5.26 10.32 8 1.38 14.27 Total 48.37 Total 44.49 6 6 45.77 12.74 6 45.77 13.24 7 9.48 12.74 Total 45.77 10 6.62 10.32 Total 61.87 7 10 230.04 10.32 6 9.48 13.70 Total 230.04 8 87.61 13.70 9 22.01 13.70 Total 119.11 ......0.00.00.........OOOOOIOOOOOOI.......-.0000...0.06.00.00.00... 8 1 8.32 11.98 8 23.40 11.98 2 13.35 11.98 Total 23.40 5 1.38 11.98 7 87.61 11.98 8 23.40 11.98 10 103.33 10.32 Total 237.40 9 7 22.01 12.59 9 23.28 13.12 9 23.28 12.59 Total 23.28 10 5.18 10.32 Total 50.47 1 Totals may not equal the sum of the parts due to rounding 174 Table 0.7. Import/Export Ouantitiaa and Prices for All Regions Under 1985 Narkat Data with Total Removal of Class I Differentials and no R0 Filtration Applied (ND) Exports Imports Region Importer Ouantity Price Exporter Ouantity Price (mil.cwt)(s/cwt) (mil.cwt.)(8/cwt) 1 1 20.58 16.15 1 20.58 16.52 10 2.02 11.78 6 2.07 16.52 Total1 22.60 Total1 22.65 2 2 26.87 14.37 2 26.87 15.11 10 2.79 11.78 5 3.8 15.11 Total 29.66 6 3.39 15.11 Total 34.07 3 3 58.14 12.62 3 58.14 13.41 10 6.37 11.78 Total 58.14 Total 64.51 4 4 50.55 11.78 4 50.55 12.47 10 15.30 11.78 Total 50.55 Total 65.85 5 2 3.80 12.94 5 44.00 14.02 5 44.00 12.94 Total 44.00 10 5.26 11.78 Total 53.06 6 1 2.07 11.78 6 45.62 12.28 2 3.39 11.78 Total 45.62 6 45.62 11.78 10 12.30 11.78 Total 63.38 7 7 118.04 11.78 7 118.04 12.50 10 122.70 11.78 Total 118.04 Total 240.75 8 8 23.32 11.78 8 23.32 11.79 10 219.07 11.78 Total 23.32 Total 242.39 9 9 23.18 11.78 9 23.18 12.32 10 28.02 11.78 Total 23.18 Total 51.20 1 Totals may not equal the sum of the parts due to rounding 175 Table 0.8. Import/Export Ouantitiaa and Prices for All Regions Under 1985 Narkat Data with Total Removal of Class I Differentials and R0 Filtration Applied at 8.90/cwt (NDR09) Exports Imports Region Importer Quantity Price Exporter Ouantity Price (mil.cwt)(8/cwt) (mil.cwt.)(8/cwt) 1 1 19.40 15.13 1 19.40 15.50 10 2.02 10.32 6 3.45 15.50 161.11 21.42 Total1 22.84 2 2 25.16 13.35 2 25.15 14.09 10 2.79 10.32 5 4.86 14.09 Total 27.94 6 4.32 14.09 Total 34.33 3 3 58.14 12.62 3 58.14 13.41 10 6.37 10.32 Total 58.14 Total 64.51 4 4 50.98 10.40 4 50.98 11.09 10 10.63 10.32 Total 50.98 Total 61.61 5 2 4.86 11.92 5 39.62 13.00 5 39.62 11.92 7 4.70 13.00 10 5.26 10.32 Total 44.32 Total 49.74 6 1 3.45 10.76 6 45.90 11.26 2 4.32 10.76 Total 45.90 6 45.90 10.76 10 6.62 10.32 Total 60.29 7 5 4.69 10.39 7 119.11 11.11 7 119.11 10.39 Total 119.11 10 97.78 10.32 Total 221.59 8 8 23.46 10.40 8 23.46 10.40 10 210.76 10.32 Total 23.46 Total 234.27 9 9 23.38 10.40 9 23.38 10.93 10 26.25 10.32 Total 23.38 Total 49.64 1 Totals may not equal the sum of the parts due to rounding 176 Table 0.9. Import/Export Quantities and Prices for All Regions Under 1985 Narkat Data with Removal of Class I Differentials and Reduced CCC Purchases (NDCCC) Exports Imports Region Importer Ouantity Price Exporter Ouantity Price (mil.cwt)(0/cwt) (mil.cwt.)(0/cwt) 1 1 18.88 14.69 1 18.88 15.05 10 2.02 11.78 6 4.05 15.05 TOtOll 20.90 161.11 22.93 2 2 25.59 13.61 2 25.59 14.35 10 2.79 11.78 5 1.73 14.35 Total 28.37 6 6.94 14.35 Total 34.26 3 3 58.14 12.62 3 58.14 13.41 10 6.37 11.78 Total 58.14 Total 64.51 4 4 50.55 11.78 4 50.55 12.48 10 15.30 11.78 Total 50.55 Total 65.86 5 2 1.73 12.37 5 44.18 13.44 5 44.18 12.37 Total 44.18 10 5.26 11.78 Total 51.17 6 1 4.05 11.78 6 45.62 12.28 2 6.94 11.78 Total 45.62 6 45.62 11.78 10 6.78 11.78 Total 63.39 7 7 118.04 11.78 7 118.04 12.50 10 122.73 11.78 Total 118.04 Total 240.77 8 8 23.32 11.78 8 23.32 11.79 10 ' 219.08 11.78 Total 23.32 Total 242.40 9 9 23.18 11.78 9 23.19 12.32 10 28.03 11.78 Total 23.19 Total 51.21 1 Totals may not equal the sum of the parts due to rounding 177 Table 0.10. Import/Export Ouantitiaa and Prices for All Regions Under 1985 Narkat Data with Class I Differential Removed, CCC Purchases Reduced and R0 Filtration Applied at 0.90/cwt (NDCCR09) Exports Imports Region Importer Ouantity Price Exporter Ouantity Price (mil.cwt)(8/cwt) (mil.cwt.)(8/cwt) 1 1 17.72 13.69 1 17.72 14.06 10 2.02 10.32 6 5.39 14.06 161611 19.74 761.11 23.11 2 2 23.91 12.61 2 23.91 13.35 10 2.79 10.32 6 2.46 13.35 Total 26.70 9 8.16 13.35 Total 34.52 3 3 58.14 12.62 3 58.14 13.41 10 6.37 10.32 Total 58.14 Total 64.51 4 4 50.93 10.56 4 50.93 11.25 10 11.18 10.32 Total 50.93 Total 62.11 5 5 43.71 11.69 5 43.81 12.76 10 5.26 10.32 7 .68 12.76 Total 48.97 Total 44.39 6 1 5.37 10.79 6 45.90 11.29 2 2.46 10.79 Total 45.90 6 45.90 10.79 10 6.62 10.32 Total 60.36 7 5 .68 10.56 7 118.99 11.28 7 118.99 10.56 Total 118.99 10 104.19 Total 223.86 8 8 23.45 10.56 8 23.45 10.57 10 211.75 10.32 Total 23.45 Total 235.19 9 2 8.16 10.56 9 23.36 11.09 9 23.36 10.56 Total 23.36 10 18.31 Total 49.82 1 Totals may not equal the sum of the parts due to rounding 178 Table 0.11. Import/Export Ouantitiaa and Prices for All Regions Under 1985 Narkat Data with 1986 Class I Differentials and no R0 Applied (D86) Exports Imports Region Importer Quantity Price Exporter Ouantity Price (mil.cwt)(s/cwt) (mil.cwt.)(s/cwt) 1 1 17.31 16.17 7 17.31 16.54 10 2.02 11.78 2 5.87 16.54 Total1 19.33 766.11 23.13 2 1 5.87 15.24 2 18.46 15.98 2 18.46 15.24 5 2.96 15.98 10 2.79 11.78 8 13.07 15.98 Total 27.12 Total 34.49 3 3 58.19 14.86 6 58.19 15.65 10 6.37 11.78 Total 58.19 Total 64.56 4 4 50.48 13.47 4 50.48 14.16 10 16.37 11.78 Total 50.48 Total 66.84 5 2 2.96 14.59 5 44.12 15.66 5 44.12 14.59 Total 44.12 10 5.26 11.78 Total 52.34 6 6 45.40 14.23 6 45.40 14.73 7 13.85 14.23 Total 45.40 10 6.62 11.78 Total 65.88 7 7 80.45 14.83 6 13.85 15.55 10 163.77 11.78 7 80.45 15.55 Total 244.22 9 23.65 15.55 Total 177.96 8 2 13.07 13.09 8 23.30 13.10 8 23.30 13.09 Total 23.30 10 207.03 11.78 Total 243.39 9 7 23.65 14.04 9 23.09 14.57 9 23.09 14.04 Total 23.09 10 5.18 11.78 . Total 51.92 1 Totals may not equal the sum of the parts due to rounding 179 Table 0.12. Import/Export Ouantitiaa and Prices for All Regions Under 1985 Narkat Data with 1986 Class I Differentials and R0 Applied at 8.90/cwt (D86R09) Exports Imports Region Importer Ouantity Price Exporter Ouantity Price (mil.cwt)(8/cwt) (mil.cwt.)(s/cwt) 1 10 19.22 11.78 8 23.45 14.79 Total1 19.22 Total1 23.45 2 10 26.90 11.78 8 34.73 14.90 Total 26.90 Total 34.73 3 10 64.52 11.78 4 10.75 15.48 Total 64.52 8 47.50 15.48 Total 58.25 4 3 10.75 13.89 4 50.36 14.58 4 50.36 13.89 Total 50.36 10 6.89 11.78 Total 68.00 5 5 33.07 14.30 5 33.07 15.37 10 18.45 11.78 8 11.13 15.37 Total 51.52 Total 44.19 6 6 45.44 14.07 6 45.44 14.57 7 13.39 14.07 Total 45.44 10 6.62 11.78 Total 65.45 7 10 244.22 11.78 6 13.39 15.12 Total 244.22 8 81.10 15.12 9 23.74 15.12 Total 118.22 3 1 23.45 13.55 3 23.25 13.56 2 34.73 13.55 166.1 23.25 3 47.50 13.55 5 11.13 13.55 7 31.10 13.55 3 23.25 13.55 10 24.72 11.73 Total 245.88 9 7 23.74 14.11 9 23.08 14.65 9 23.08 14.11 Total 23.08 10 5.18 11.78 Total 51.99 1 Totals may not equal the sum of the parts due to rounding 180 Table 0.13. Import/Export Ouantitiaa and Prices for All Regions Under 1985 Narkat Date with Transportation Costs Doubled and no R0 Filtration Applied (T02) Exports Imports Region Importer Ouantity Price Exporter Quantity Price (mil.cwt)(0/cwt) (mil.cwt.)(s/cwt) 1 1 19.66 18.63 1 19.66 19.17 10 2.02 11.78 2 3.12 19.17 Total1 21.63 761.11 22.73 2 1 3.12 16.32 2 22.75 17.42 2 22.75 16.32 5 11.42 17.42 10 2.79 11.78 Total 34.17 Total 28.65 3 3 58.06 14.83 3 58.06 16.01 10 6.37 11.78 Total 58.06 Total 64.43 4 4 50.45 13.22 4 50.45 14.26 10 15.71 11.78 Total 50.45 Total 66.16 5 2 11.42 14.72 5 36.01 16.32 5 36.01 14.72 7 7.93 16.32 10 5.26 11.78 Total 43.94 Total 52.69 6 6 45.53 13.46 6 45.53 14.21 10 18.25 11.78 Total 45.53 Total 63.81 7 5 7.93 14.57 7 117.89 15.65 7 117.89 14.57 Total 117.89 10 115.50 11.78 Total 241.32 8 8 23.28 12.98 8 23.28 13.24 10 219.52 11.78 Total 23.28 Total 242.81 9 9 23.14 13.38 9 23.14 14.18 10 28.12 11.78 Total 23.14 Total 51.26 1 Totals may not equal the sum of the parts due to rounding 181 Table C.14. Import/Export Ouantitiaa and Prices for All Regions Under 1985 Narkat Data with Transportation Costs Increased Fifty Percent and R0 Filtration Applied at 8.90/cwt (TC2RO9) Exports Imports Region Importer Ouantity Price Exporter Ouantity Price (mil.cwt)(s/cwt) (mil.cwt.)(0/cwt) 1 1 17.57 16.44 1 17.57 16.99 10 2.02 11.78 5 2.32 16.99 Total1 19.59 3 3.123 16.99 Total 23.12 2 2 24.21 15.15 2 24.21 16.26 10 2.79 11.78 4 8.92 16.26 Total 27.00 8 1.30 16.26 Total 34.43 3 3 58.06 14.83 3 58.06 16.01 10 6.37 11.78 Total 58.06 Total 64.43 4 2 8.92 13.26 4 50.44 14.29 4 50.44 13.26 Total 50.44 10 6.89 11.78 Total 66.25 5 1 2.32 14.34 5 44.04 15.94 5 44.04 14.34 Total 44.04 10 5.26 11.78 Total 51.61 6 6 45.37 14.10 6 45.37 14.85 7 13.55 14.10 Total 45.37 10 6.62 11.78 Total 65.54 7 7 80.89 14.57 6 13.55 15.65 10 160.44 11.78 7 80.89 15.65 Total 241.33 9 23.45 15.65 Total 117.89 8 1 3.23 12.98 8 23.28 13.24 2 1.30 12.98 Total 23.28 8 23.28 12.98 10 214.99 11.78 Total 242.81 9 7 23.45 13.83 9 23.08 14.63 9 23.08 13.83 Total 23.08 10 5.18 11.78 Total 51.71 1 Totals may not equal the sum of the parts due to rounding 182 Table 0.15. Import/Export Ouantitiaa and Prices for All Regions Under September 1985 Narkat Data with no R0 Filtration Applied (SEPT) Exports Imports Region Importer Ouantity Price Exporter Quantity Price (mil.cwt)(s/cwt) (mil.cwt.)(s/cwt) 1 1 1336.19 16.13 1 1336.19 16.50 10 151.58 11.78 6 717.19 16.50 763611 1487.76 761.11 2053.37 2 2 1893.81 14.90 2 1893.81 15.64 10 212.97 11.78 6 244.08 15.64 Total 2106.78 9 867.55 15.64 Total 3005.44 3 3 5020.39 15.02 3 5020.39 15.80 10 507.86 11.78 Total 5020.39 Total 5528.25 4 4 4677.66 12.54 4 4677.66 13.23 10 1005.51 11.78 Total 4677.66 Total 5683.17 5 5 3962.73 14.21 5 3962.73 15.28 10 436.19 11.78 9 15.68 15.28 Total 4398.92 Total 3978.41 6 1 717.19 13.12 6 3886.22 13.62 2 244.08 13.12 Total 3886.22 6 3886.22 13.12 7 16.38 13.12 10 554.15 11.78 Total 5418.02 7 7 9819.00 13.89 6 16.38 14.31 10 9947.47 11.78 7 9819.00 14.61 Total 19766.47 Total 9835.38 8 8 1910.32 12.30 8 1910.32 12.31 10 17609.27 11.78 Total 1910.32 Total 19519.60 9 2 867.55 12.70 9 1926.36 13.24 5 15.68 12.70 Total 1926.36 9 1926.36 12.70 10 1512.84 11.78 Total 4322.42 1 Totals may not equal the sum of the parts due to rounding 183 Table 0.16. Import/Export Quantities and Prices for All Regions Under September 1985 Narkat Data with R0 Filtration Applied at 8.30/cwt (SRO3) Exports Imports Region Importer Ouantity Price Exporter Ouantity Price (mil.cwt)(8/cwt) (mi1.cwt.)(8/cwt) 1 10 1336.73 11.78 6 873.78 14.10 Total1 1336.73 3 1213.54 14.10 Total1 2037.33 2 10 1954.54 11.78 8 3042.26 13.79 Total 1954.54 Total 3042.26 3 3 4349.02 13.36 3 4319.02 14.15 10 507.86 11.78 4 726.30 14.15 Total 4856.88 Total 5075.32 4 3 726.30 12.55 8 4679.28 13.17 10 4958.16 11.78 Total 4679.28 Total 5684.50 5 5 3719.34 13.22 5 3719.34 14.06 10 436.19 11.78 8 290.17 14.06 Total 4155.53 Total 4009.50 6 1 873.78 12.79 6 3893.52 13.29 6 3893.52 12.79 Total 3893.52 7 18.70 12.79 10 554.15 11.78 Total 5340.15 7 7 9882.00 13.90 6 18.70 13.11 10 9889.47 11.78 7 9882.00 14.56 Total 19771.47 Total 9900.70 8 1 1213.54 12.31 8 1910.28 12.31 2 3042.26 12.31 Total 1910.28 4 4679.28 12.31 5 290.17 12.31 8 1910.28 12.31 10 8386.42 11.78 Total 19524.95 9 9 1926.30 12.71 9 1926.30 13.24 10 2396.58 11.78 Total 1926.30 Total 4322.88 1 Totals may not equal the sum of the parts due to rounding 184 Table 0.17. Import/Export Ouantitiaa and Prices for All Regions Under September 1985 Narkat Data with R0 Applied at 8.90/cwt (SR09) Exports Imports Region Importer Ouantity Price Exporter Ouantity Price (mil.cwt)(I/cwt) (mil.cwt.)(8/cwt) 1 1 1198.83 14.29 1 1198.83 14.65 10 151.58 11.78 6 880.64 14.65 Total1 1350.41 763611 2079.47 2 2 1752.49 13.65 2 1752.49 14.39 10 212.97 11.78 8 1277.81 14.39 Total 1965.46 Total 3030.30 3 3 4621.51 14.04 3 4621.51 14.82 10 507.86 11.78 4 431.52 14.82 Total 5129.37 Total 5053.02 4 3 431.52 12.62 4 4675.63 13.31 4 4675.63 12.62 Total 4675.63 10 594.80 11.78 Total 5701.94 5 5 3808.49 13.58 5 3808.49 14.66 10 426.19 11.78 8 185.82 14.66 Total 4244.69 Total 3994.32 6 1 880.64 12.75 6 3894.53 13.25 6 3894.53 13.75 Total 3894.53 10 554.15 11.78 Total 5329.32 7 7 9882.00 13.90 7 9882.00 14.61 10 9887.93 11.78 Total 9882.00 Total 19769.93 8 2 1277.81 12.31 8 1910.29 12.31 5 185.82 12.31 Total 1910.29 8 1910.29 12.31 10 16147.29 11.78 Total 19521.22 9 9 1926.32 12.71 9 1926.31 13.24 10 2396.42 11.78 Total 1926.31 Total 4322.73 1 Totals may not equal the sum of the parts due to rounding 185 Table 0.18. Import/Export Ouantitiaa and Prices for All Regions Under September 1985 Narkat Data with R0 Filtration Applied at 81.75/cwt (SR0175) Exports Imports Region Importer Ouantity Price Exporter Ouantity Price (mil.cwt)(0/cwt) (mil.cwt.)(s/cwt) 1 1 1262.02 15.13 1 1262.02 15.50 10 151.58 11.78 6 805.44 15.50 761.11 1413.60 763311 2067.47 2 2 1848.74 14.50 2 1848.74 15.24 10 212.97 11.78 8 1164.63 15.24 Total 2061.71 Total 3013.37 3 3 4936.65 14.81 3 4936.65 15.60 10 507.86 11.78 4 90.58 15.60 Total 5444.51 Total 5027.24 4 3 90.58 12.54 4 4677.62 13.24 4 4677.62 12.54 Total 4677.62 10 915.28 11.78 Total 5683.49 5 5 3963.04 14.21 5 3963.04 15.28 10 436.19 11.78 9 15.33 15.28 Total 4399.23 Total 3978.37 6 1 805.44 12.74 6 3894.59 13.24 6 3894.59 12.74 Total 3894.59 10 628.63 11.78 Total 5328.66 7 - 7 9882.00 13.89 7 9882.00 14.61 10 9885.70 11.78 Total 9882.00 Total 19767.70 8 2 1164.63 12.30 8 1910.31 12.31 8 1910.31 12.30 Total 1910.31 10 16445.24 11.78 Total 19520.17 9 5 15.33 12.70 9 1926.34 13.24 9 1926.34 12.70 Total 1926.34 10 2380.86 11.78 Total 4322.53 1 Totals may not equal the sum of the parts due to rounding 186 Table 0.19. Import/Export Ouantitiaa and Prices for All Regions Under Nuy’s New Supply Elaaticitiaa with Regions 9’s Elasticity Set Equal to .00, ggtgxig_pg;ihgg_(NUY00) Exports Imports Region Importer Ouantity Price Exporter Quantity Price (mil.cwt)(8/cwt) (mil.cwt.)(s/cwt) 1 1 18.14 17.05 1 18.14 17.42 10 2.02 10.32 2 4.91 17.42 T0t811 20.16 Total1 23.05 2 1 4.91 15.69 2 20.16 16.43 2 20.16 15.69 4 10.69 16.43 10 2.79 10.32 5 3.54 16.43 Total 27.85 Total 34.39 3 3 58.16 14.94 3 58.16 15.73 10 6.37 10.32 Total 58.16 Total 64.53 4 2 10.69 13.28 4 706.08 13.97 4 50.53 13.28 Total 706.08 10 6.89 10.32 Total 68.11 5 2 3.54 14.81 5 44.06 15.88 5 44.06 14.81 Total 44.06 10 5.26 10.32 Total 52.86 6 6 45.45 14.06 6 45.45 14.55 7 13.84 14.06 Total 45.45 10 6.62 10.32 Total 65.90 7 7 80.96 14.57 6 13.84 15.29 10 162.29 10.32 7 80.96 15.29 Total 243.25 9 23.32 15.29 Total 118.12 8 8 23.31 12.98 8 23.31 12.99 10 223.89 10.32 Total 23.31 Total 247.20 9 7 23.32 13.67 9 23.14 14.20 9 23.14 13.67 Total 23.14 10 5.18 10.32 Total 51.76 1 Totals may not equal the sum of the parts due to rounding 187 Table 0.20. Import/Export Ouantitiaa and Prices for All Regions Under Nuy’s New Supply Elasticities with Regions 9’s Elasticity Set Equal to .05, ggtgzig_pgzipgg_(HUY05) Exports Imports Region Importer Ouantity Price Exporter Quantity Price (mil.cwt)(8/cwt) (mil.cwt.)(8/cwt) 1 1 18.14 17.05 1 18.14 17.42 10 2.02 10.32 2 4.91 17.42 163611 20.16 Total1 23.05 2 1 4.91 15.69 2 20.16 16.43 2 20.16 15.69 4 10.69 16.43 10 2.79 10.32 5 3.54 16.43 Total 27.85 Total 34.39 3 3 58.16 14.94 3 58.16 15.73 10 6.37 10.32 Total 58.16 Total 64.53 4 2 10.69 13.28 4 706.08 13.97 4 50.53 13.28 Total 706.08 10 6.89 10.32 Total 68.11 5 2 3.54 14.81 5 44.06 15.88 5 44.06 14.81 Total 44.06 10 5.26 10.32 Total 52.86 6 6 45.45 14.06 6 45.45 14.55 7 13.84 14.06 Total 45.45 10 6.62 10.32 Total 65.90 7 7 80.96 14.57 6 13.84 15.29 10 162.29 10.32 7 80.96 15.29 Total 243.25 9 23.32 15.29 Total 118.12 8 8 23.31 12.98 8 23.31 12.99 10 223.89 10.32 Total 23.31 Total 247.20 9 7 23.45 13.67 9 23.14 14.20 9 23.14 13.67 Total 23.14 10 5.18 10.32 Total 51.77 1 Totals may not equal the sum of the parts due to rounding 188 Table 0.21. Import/Export Ouantitiaa and Prices for All Regions Under Huy’s New Supply Elasticities with Regions 9’s Elasticity Set Equal to .10, ggtggig_pggihgg,(NUY10) Exports Imports Region Importer Ouantity Price Exporter Ouantity Prica (mil.cwt)(I/cwt) (mil.cwt.)(s/cwt) 1 1 18.14 17.05 1 18.14 17.42 10 2.02 10.32 2 4.91 17.42 761.11 20.16 761.11 23.05 2 1 4.91 15.69 2 20.16 16.43 2 20.16 15.69 4 10.69 16.43 10 2.79 10.32 5 3.54 16.43 Total 27.85 Total 34.39 3 3 58.16 14.94 3 58.16 15.73 10 6.37 10.32 Total 58.16 Total 64.53 4 2 10.69 13.28 4 706.08 13.97 4 50.53 13.28 Total 706.08 10 6.89 10.32 Total 68.11 5 2 3.54 14.81 5 44.06 15.88 5 44.06 14.81 Total 44.06 10 5.26 10.32 Total 52.86 6 6 45.45 14.06 6 45.45 14.55 7 13.84 14.06 Total 45.45 10 6.62 10.32 Total 65.90 7 7 80.96 14.57 6 13.84 15.29 10 162.29 10.32 7 80.96 15.29 Total 243.25 9 23.32 15.29 Total 118.12 8 8 23.31 12.98 8 23.31 12.99 10 223.89 10.32 Total 23.31 Total 247.20 9 7 23.32 13.67 9 23.14 14.20 9 23.14 13.67 Total 23.14 10 5.18 10.32 Total 51.64 1 Totals may not equal the sum of the parts due to rounding 189 Table 0.22. Import/Export Quantities and Prices for All Regions Under Nuy’s New Supply Elasticities with Regions 9’s Elasticity Set Equal to .15, ggtggig_pg;ipngq(NUY15) Exports Imports Region Importer Ouantity Price Exporter Ouantity Price (mil.cwt)(8/cwt) (mil.cwt.)(8/cwt) 1 1 18.14 17.05 1 18.14 17.42 10 2.02 10.32 2 4.91 17.42 Total1 20.16 161.11 23.05 2 1 4.91 15.69 2 20.16 16.43 2 20.16 15.69 4 10.69 16.43 10 2.79 10.32 5 3.54 16.43 Total 27.85 Total 34.39 3 3 58.16 14.94 3 58.16 15.73 10 6.37 10.32 Total 58.16 Total 64.53 4 2 10.69 13.28 4 706.08 13.97 4 50.53 13.28 Total 706.08 10 6.89 10.32 Total 68.11 5 2 3.54 14.81 5 44.06 15.88 5 44.06 14.81 Total 44.06 10 5.26 10.32 Total 52.86 6 6 45.45 14.06 6 45.45 14.55 7 13.84 14.06 Total 45.45 10 6.62 10.32 Total 65.90 7 7 80.96 14.57 6 13.84 15.29 10 162.29 10.32 7 80.96 15.29 Total 243.25 9 23.32 15.29 Total 118.12 8 8 23.31 12.98 8 23.31 12.99 10 223.89 10.32 Total 23.31 Total 247.20 9 7 23.32 13.67 9 23.14 14.20 9 23.14 13.67 Total 23.14 10 5.18 10.32 Total 51.65 1 Totals may not equal the sum of the parts due to rounding BIBLIOGRAPHY BIBLIOGRAPHY Babb, E.N., D.E. Banker, 0. Goldman, D.R. Nartella, and J-E- PrItt- W W- WIIt Lafayottoz Purdue University Press, Agricultural Experiment Station Bul- latin No.157, April 1977. Babb, E.N., D.E. Banker and G.L. Nelson. W; W- Indium Agricultural Station Bulletin No.146, November 1976. Babb, E.N. and Arlo J. Ninden. ”A Total Narkating System for U.S. Dairy Cooperatives." Qaazlaan_1aa;aa1_afi, WW- 2(19'72) 3271-274- Beck, R.L., and J.D. Goodin. "Optimal Number and Location of Nanufactured Nilk Plants to Ninimize Narkating COItI - " WWW- 12(1980):103-108. Besik, F. ”Reverse Osmosis in Processing Whey." Qanaflian, W- July 1971. 9932-35- Blakeslee, L.L., E.0. Heady, and C.F. Framingham. flank: W- AI“: Iowa Stat-I University Press, 1973. Blakley, L.V. "Why Not a Flat Class I Price of Nilk ”ItionwidI- " WW flazkaalna_§anfia;anaa, Nadiaon: University of Wisconsin Press, April 1967, pp.60-61. Boobs. W111“- T- W 1 d - W !,'--,- ‘l._‘ 0111- I! Ih£233h_12223 Blacksburg: Virginia Polytechnic Institute and State University Press, Research Division Bulletin No. 120, December 1976. Breasler, Raymond Jr. and Richard A. King. flagkaaa‘_£;;aaa, W- Rad-19h. NC: John 1011-7 Ind Sons Inc., 1978, pp.86-92. 190 191 Buccola, Steven T. and N.C. Conner. Eaaaaala;_§1j1alaaagla A n ' -l,'. x 1 x-nu - " WWW- Blacksburg: Virginia Polytechnic Institute and State University Press, Research Division Bulletin No. 149, 1979. Buxton, Boyd N., Tom NcGuckin, Roger Selley and Gayle Willett. WWW iaan_. U. S. Department of Agriculture, Economics Research Service Agricultural Economics Report No. 528, February 1985. Chane. CJI- W W WM- Radium = University of Wisconsin, Ph.D. Dissertation, 1972. Cilley, Bradly and Leo V. Blakley. Baalanal_£ziaa_§1aa_ WW StillwIt-Ir 0K= Oklahou State University Press, Agricultural Experiment Station Bulletin No. 4329, 1985. Cox, 0.0. and I.A. Langdon ”Economic Evaluation of Reverse Osmosis for Reduction in Nilk Transport Costs.“ Iha_ . September 1985, pp.113-117. Dixon, B.D. "Dairy Products Prepared from Reverse Osmosis Concentrate--Narket Nilk Products, Butter, Skim Nilk Poder Ind Yoehurto” WW nalaax, September 1985, pp.91-95. Dunkley, W.L. "Ultra-Filtration and Reverse Osmosis in Whey WWII-ing - " WWW- June 1971, pp.47-52. Fall-rt. R-F- WWW- U.S. Department of Agriculture, Economic Research Service, Agricultural Information Bulletin No. 474, September 1984. Fallert, Richard F. and Boyd N. Buxton. Al;a;na;ixa_figaaina, W W- U-S- Mutant of Agriculture, Economics, Statistics, and Cooperative Ser- vice, Agricultural Economics Report No. 401, Hay 1978. Fenton-Nay, R.I., 0.6. Hill Jr., 0.". Amundson, and P.D. Auclair. "The Use of Pressure Driven Nembrane Processes in the Dairy Industry.“ Al§n§_§xaaaa;aa_§agiaa. Nadison: University of Wisconsin, Series No. 120, Volume 68. 192 Fenton-Hay, R.I., 0.6. Hill Jr., 0.H. Amundson, H.H. Lopez, and P.D. Auclair. “Concentration and Fractionation of Skimmilk by Reverse Osmosis and Ultrefiltration." laagaal, a£_na;;x_§aianaa, 55(1972):1561-1566. GIorQI. P-S- Ind G-A- Kinq- W 1239, Berkeley: University of California Press, Agricul- tural Experiment Station, Giannini Foundation Honograph No. 26, Harch 1971. Goldsmith, R.L., Clyde H. Amundson, Bernard S. Horton, Steven R. Tannenbaum. “Recovery of Cheese Whey Proteins Through Ultrafiltration.“ Washington D.C., Conference Paper, August 1970. cu“. LL and LL JIcobIon- W W 219211812 Hadison: University of Wisconsin Press, Agricultural Economics Research Report No. R-2503, June 1973. Griffith, David L. “Hembrane Technology." Eagnaa‘aa_ana, W- Edit-Id by "-8- Groig. 1984' ppa239-2‘1e Hahn, David E. ”Cost of Bulk Nilk Assembly.“ 212911§1821_Q£. fl§;12§_ga£hanaa, edited by Andrew H. Novakovic, Ithica: Cornell University Press, Northeast Regional Research Publication No. 39, 1983. HIIIbIrs H-C- Ind R-F- FIUIrt- W . State College: Penn- sylvania State University Press, Agricultural Experiment Station Bulletin No. 805, January 1976. Hallberg, H.C., D.E. Hahn, R.W. Stammer, G.J. Elterich, 64» FifI- WW WWW gaaaag, State College: Pennsylvania State University Press, Agricultural Experiment Station Bulletin No. 818, Hay 1978. Hammond. Jere-I U. W- 31.9.1.1: University of Hinnesota Press, Department of Agricul- tural and Applied Economics Staff Paper No. 74-12, July 1974. Hammond, Jerome W., Boyd H. Buxton and Cameron S. Thraen. . Hinneapolis: Univer- sity of Hinnesota Press, Agricultural Experiment Station Bulletin No. 529, 1979. 193 Holland. ForrIIt 0- WW . U.S. Department of Agri- culture, Economics Research Service, Staff Report No. AGES850514, October 1985. Holland, F.D., and J.E. Pratt. flaaa;_fi_£QfiIfiAfl_fi;aazaa_, . . '._ ., I ‘ 1' II II x '1 W W- "IIt Lei-nth: PurdUI University Press, Agricultural Experiment Station Bulletin No. 296, November 1980. Holland, Forrest D. and Jerry A. Sharplea. H2214_flh11$. ' . West Lafayette: Purdue University Press, Department of Agricultural Econ- omics Staff Paper No. 84-20, November 1984. Horton, Bernard S. “Whey Processing By Sanitary CIP Hembrane Separation Equipment." Aaa;19an_£a;;x_flaxiag, April 1973, pp.45-50. Huv. Birgit- WW -- c r . University of Delaware, Department of Food and Resource Economics, Unpublished Haster's Thesis, 1986. Jun. Ed - WW :a;iaa_fiaahaaa, U.S. Department of Agriculture, Economics and Statistics Service, Staff Report No. 801201, December 1980. Johnston, N. "Improvements in Whey Concentration by Reverse Osmosis." Whitchurch, U.K.: Paterson Candy Inter- national, Undated Hemeograph. Kloth, D.W. and L.V. Blakley. "Optimal Dairy Plant Location with Economics of Size and Narkat Share." Aaa;;aan_jaa;;_ W- 53(1971H461-466- Kuhn, H.W. and J.G. Hackinnon. "Sandwich Hethod for Finding Fix-d Pointh" W W- 17<1975>=139-204. Loeb,S. and SouriraJan. Title Unavailable. Qagaagaa_ W- 38(1962H117-32- Lough. H-"- WW- "-3- Department of Agricultural, Economics Research Service, August 1977. HacKinnon, James G. ”An Algorithm for the Generalized Transportation Problem.” MW- “1979:1145-64- 194 Harshall, 5.0. "The Reverse Osmosis Process." 33.45.231.133 W— SIptIIbIr 1985. pp-86-87- _. ”Reverse Osmosis Plant Operation. ” 1311231110 W SIptIIbIr 1983. pp 88- 89- NcBridI. 6- and mm Boynton- “Mammalian Q2IE_5LIH££H£!_ln_L2!!£_E1£hlS£nm EOIt Lansing: Nichigan State University Press, Agricultural Experiment Station, December 1976. HcGIrrY. H-J- MW- "hint-ity of Wisconsin, Unpublished Ph.D. Dissertation, 1968. NcKenna, Brian H. “Reverse Osmosis: A New Concept in Concentration for the Dairy Industry.“ Qa1;x_1naaaaziaa. 35(1970):755-757. Nerson, Richard L., and L. Frank Ginnette. “Improved Processing of Foods by Reverse Osmosis." U. S. Department of Agriculture, Agricultural Research Service, Hemeo- graphed Paper, 1970. Honda. Rorbort H- W W- U-S- DIP-rtI-Int of Agriculture, Economic Research Service, Hay 1979. Hohr, Paula. “Western Dairies Square-Off Against the Hidwest." Ea;a_laazna1, Harch 1986, pp.12-19. Hoore, R., S. Elassar, and B.V. Lessley. - W . College Park: Uni- versity of Naryland Press, Agricultural Experiment Sta- tion Bulletin No. 796, 1972. National Hilk Producers Federation. “1984 Dairy Producer Highlights.“ Arlington VA: NHPF, 1985. Matt. Sherril 8- W W- But Lansing: Nichigan State University Press, Department of Agricultural Economics Staff Paper No. 85-45, 1985. ngnn, East Lansing: Nichigan State University Press, Department of Agricultural Economics Staff Paper No. 85- 50, 1985. Novakovic. AndrIw- W W- Ithica: Cornell University Press, Agricultural Experiment Station Research Report No. 82-32, October 1982. 195 Novakovic, Andrew and Richard Aplin, Eh3&_ll_£h1_21££££1821 w v h - gaaa;Qa_. Ithica: Cornell University Press, Agricultural Experiment Station Economics Research Report No. 81-19, September 1981. Novakovic, A.H., E.N. Babb, D.R. Nertelle, and J.E. Pratt. Ea;;ax_§;aa1aaaz, Ithica: Cornell University Press, Agricultural Experiment Station Research Report No.80-21, 1980. Novakovic, A.H. and J.E. Pratt. fledglins_ths_§2:iisl WW- Ithica: Cornell University Press, Agricultural Experiment Station Staff Paper No. 86-5, Harch 1986. Oamonica, Inc. Untitled Brochure. Hinnetonka, HN. Undated. Ottem, Hike. Research and Development Economist. Ridgeview Enterprises, La Crosse, WI. Telephone Conversation Sepember 9, 1986. Pratt. Jon-I E- W W- Nichigan State University, Unpublished Ph.D. Dissertation, 1986. Pratt, J.E., A.H. Novakovic, G. J. Elterich, D. E. Hahn, B.J. Smith, and G. K. Criner. “An Analysis of the Spatial Organization of the Northeast Dairy Induatry. " Saagah3__ Aa;;aal§a;a,. Ithica: Cornell University Press, Agricul- tural Experiment Station Report No. 32, 1986. Pratt, James E., Andrew H. Novakovic, and David Jensen. "The Northeast Dairy Sector Simulator.“ E£2£££§1821_2£, W . Edited by Andrew H. Novak- ovic. Ithica: Cornell University Press, Northeast Regional Research Publication No. 39, 1983. Rand HcNally. ' l l . Chicago: Rand HcNally, 1981. Randolph. Tho-II F- W W Eaaa_§aaaaay, Cornell University, Unpublished Master’s Thesis, June 1986. Reid. c.1-:. and Ed. Broton- W W 1267. State College: Pennsylvania State University Press, Agricultural Experiment Station Bulletin No. 783, August 1972 196 Riley, J.B. and L.V. Blakley. "Impacts of Alternative Class I Pricing Systems of Fluid Nilk Prices." Aaa;;aan_laa;aa1 cf_Asrisnlinznl_fiscnosicn. 57(1975>:67-73. .______.- Eauilibriua_Ia2a2t_2f_Altar011121.2zisins_fiolisieg. ana_§tr2si2zal_Qh2nss1_in_th:_Elni§_£il£_ln§2:trxi. 1222;121fi. Stillwater OK: Oklahoma State University Press, Agricultural Experiment Station Research Report No. p-733, Hay 1976. Roberts, Tayna. “Rethinking Reconatituted Nilk Reg- ulations.” flaaiana1_£aaa_fiagiag. U.S. Department of Agriculture, Economics, Statistics and Cooperative Service, NFR-lo, 1980. Selathe, Larry E., J. Hichael Price, and Kenneth E. Gadson. “The Food and Agricultural Policy Simulator: The DIirv-Soctor SubIodI1-“ Aszisnltntnl_§sononisg_fisz. Illifihr 34(3), July 1982, pp.1-14. Samuelson, P.A. ”Spatial Price Equilibrium and Linear Programming." Aaagiaan_fiaanaaia_fiayiag, 42(1952):283-303. Schmitz, A. in_12§9, University of Wisconsin, Unpublished Ph.D. Dissertation, 1968. Seton, Nora and Ronald Knutson. “Where is the Lowest Cost Milk Producod?” flos£§:_nsitzsnn. SoptIIbIr 1985. pp.1173,1228. Stewart. ChorIII- Ec2n2ais_E1s1ibiliix_cf_s_§ssll_fln:lhe;. EAza_nilt_zzos:snins_filsnt.- U 8- DIpIrtIInt of Agricul- ture, Narkat Research Report No. 1087, 1978. Takayama, T. “Spatial Equilibrium and Ouadratic Program I1n9-“ l2ntnal_21_Esza_Econosissy 46=67-93- ._______- 52stial_1nd_Iss22r1l_2zic:.nn§.Allocstien.flo§elg. Amsterdam: North Holland Press, 1971. Tokovo-o. T- and H HoIhinotO- anssis_nsrkztzgzisntgs. 22rlc_E92d_EI212sti2n_and_Elsnnins_nessls_ss§_lhsir. 7 - 7 W - tian, Urbana: University of Illinois Press, World Food Proaection Pronect Report No. 2, 1976. Takayama, T. and G.G. Judge. "Equilibrium Among Spatially Separated Harkets: A Reformulation.“ Eganaaaagiga. 32(1964):510-524. Takayama, T. and C.L. Liu. "Projections of International Trade in Farm Products I: Wheat." I;11naia_fia£ignltazal, Eganaaiaa, 15(1975):1-7. 197 Tho-u. VLA- Ind R.K- DIHIVIn- W W Qagaliaa, Clemson, 80: Clemson University Press, South Carolina Agricultural Experiment Station Bulletin No. 603, October 1977. 0-8- DIP-rtnnt of Agriculture. W W- Scone-1“ RIM-arch Service, Staff Report No. AGES840121, January 1984. fi-WWW 7 e m . pre- pared for the Committee on Agriculture, Nutrition, and Forestry, U.S. Senate, June 1979. _______, “Dairy Narkat News.“ Agricultural Narkating Service, Volume 52, Report 12, Narch 1985. Agricultural Narkating Service, Narch 1986. figaaagaian‘_12§2. U.S. Department of Agriculture, Economic Research Service, ECIFS 2-3, December 1983. Egaaaaaiaa‘_12§a, U.S. Department of Agriculture, Economic Research Service, ECIFS 3-1, July 1984. __-WWW_Q£ EIQQH£11981_12§19 U.S. Department of Agriculture, Economic Research Service, ECIFS 4-1, September 1985. _-mwmmJWmm Saaaang. Agricultural Narkating Service, Statistical Bulletin No. 745, July 1986. _WMWLW lgga, Agricultural Harketing Service, FHOS No. 309, Harch 1986. m-WMW 125:, Agricultural Narkating Service, FHOS No. 312, Hay 1986. _-Wummmmm ifiafi. Agricultural Narkating Service, FHOS No. 315, August 1986. . "Firm Enterprise Data System: Dairy Budgets, 1982". Provided under a cover letter by Timothy J. Ulrich, unpublished data, 1983. 198 _-MMHWW ana_Laaa1_§;aaaa, Economic Research Service, Staff Report No. AGESB60404, April 1986. United States Department of Commerce. 12§Q_§anaaa_a£ WW- Bur-nu of Con-us. P080-1-A1, p.1-22, Figure 19. ‘ __-Wunms_mmimm Qaazaaaagiaalaa, Bureau of Census, P080-1-B, pp.1-9, Figure 3. Bureau of the Census, A082-SS-1, October 1985, pp.115 116. United States Public Health Service. ”Grade A Pasteurization Nilk Ordinance.” USPHS No. 229, 1965, p.2. Versteeg, 0. "Process and Ouality Control Aspects of Reverse Osmosis Concentration of Hilk.“ 19.43.211.133 W- Soot-loot 1985. pp-102-107- "Commercial Reverse Osmosis Plants.” __ W3. W- SIptIIbIr 1985. 139-118421- Warner, D.L. W Princeton University, Unpublished Ph.D Dissertation, 1979. Whipple, Glen D. “An Analysis of Reconstituted Fluid Hilk Pricing Pelican" MW 9213:» 65(1983):214-24.